Rheumatology 2021;60:2615–2628 Rheumatology doi:10.1093/rheumatology/keab166 Advance Access publication 18 February 2021 Systematic review and meta analysis A systematic review and meta-analysis to inform cancer screening guidelines in idiopathic inflammatory myopathies Alexander G. S. Oldroyd 1,2,3,4, Andrew B. Allard5, Jeffrey P. Callen6, 1,2,4 Hector Chinoy , Lorinda Chung7,8, David Fiorentino9, Michael D. George10,11, Patrick Gordon12, Kate Kolstad9, Drew J. B. Kurtzman13, Pedro M. Machado 14,15,16 , Neil J. McHugh 17 , Anna Postolova , Albert Selva-O’Callaghan , Jens Schmidt20, 18 19 5,20 Sarah Tansley , Ruth Ann Vleugels21,22, Victoria P. Werth23,24 and 25 Rohit Aggarwal CLIN IC AL S C IE N C E Abstract Objectives. To identify clinical factors associated with cancer risk in the idiopathic inflammatory myo- pathies (IIMs) and to systematically review the existing evidence related to cancer screening. Methods. A systematic literature search was carried out on Medline, Embase and Scopus. Cancer risk within the IIM population (i.e. not compared with the general population) was expressed as risk ratios (RR) for binary variables and weighted mean differences (WMD) for continuous variables. Evidence relating to cancer screening practices in the IIMs were synthesized via narrative review. Results. Sixty-nine studies were included in the meta-analysis. DM subtype (RR 2.21), older age (WMD 11.19), male sex (RR 1.53), dysphagia (RR 2.09), cutaneous ulceration (RR 2.73) and anti-transcriptional intermediary fac- tor-1 gamma positivity (RR 4.66) were identified as being associated with significantly increased risk of cancer. PM (RR 0.49) and clinically amyopathic DM (RR 0.44) subtypes, Raynaud’s phenomenon (RR 0.61), interstitial lung disease (RR 0.49), very high serum creatine kinase (WMD 1189.96) or lactate dehydrogenase (WMD 336.52) levels, and anti-Jo1 (RR 0.45) or anti-EJ (RR 0.17) positivity were identified as being associated with sig- nificantly reduced risk of cancer. Nine studies relating to IIM-specific cancer screening were included. CT scan- ning of the thorax, abdomen and pelvis appeared to be effective in identifying underlying asymptomatic cancers. Conclusion. Cancer risk factors should be evaluated in patients with IIM for risk stratification. Screening evidence is limited but CT scanning could be useful. Prospective studies and consensus guidelines are needed to establish cancer screening strategies in IIM patients. 1 National Institute for Health Research Manchester Biomedical Research Centre, Manchester University NHS Foundation Trust, 2Centre for Musculoskeletal Research, University of Manchester, Manchester Academic Health Science Centre, 3Centre for Epidemiology Versus Arthritis, University of Manchester, Manchester, 4Department of Rheumatology, Salford Royal NHS Foundation Trust, Salford, 5Royal National Hospital for Rheumatic Diseases, Royal United Hospitals Bath NHS Foundation Trust, Bath, UK, 6Division of Medicine, Department of Medicine, University of Louisville, Louisville, KY, 7Division of Immunology and Rheumatology, Department of Medicine and Dermatology, Stanford University, Stanford, 8Palo Alto Health Care System, Palo Alto, 9Department of Dermatology, School of Medicine, Stanford University, Stanford, CA, 10Division of Rheumatology, 11Division of Epidemiology, University of Pennsylvania, Philadelphia, PA, USA, 12Department of Rheumatology, King’s College Hospital NHS Foundation Trust, London, UK, 13Department of Dermatology, Wright State University, Dayton, OH, USA, 14Centre for Rheumatology and Department of Neuromuscular Diseases, University College London, 15National Institute for Health Research University College London Hospitals Biomedical Research Centre, University College London Hospitals NHS Trust, 16Department of Rheumatology, Northwick Park Hospital, London North West University Healthcare NHS Trust, London, 17Department of Pharmacy and Pharmacology, University of Bath, Bath, UK, 18Department of Immunology and Rheumatology, Stanford Health Care, Stanford, CA, USA, 19Systemic Autoimmune Unity, Vall D’Hebron General Hospital, Universitat Autonoma de Barcelona, Barcelona, Spain, 20Department of Neurology, Neuromuscular Centre, Göttingen, Germany, 21Harvard Medical School, 22Department of Dermatology, Brigham and Women’s Hospital, Boston, MA, 23Department of Dermatology, University of Pennsylvania, 24Division of Dermatology, Corporal Michael J. Crescenz VA Medical Center, Philadelphia and 25Division of Rheumatology and Clinical Immunology, Department of Medicine, University of Pittsburgh, Pittsburgh, PA, USA Submitted 4 December 2020; accepted 11 February 2021 Correspondence to: Rohit Aggarwal, Division of Rheumatology and Clinical Immunology, Arthritis and Autoimmunity Center (Falk), UPMC Myositis Center, University of Pittsburgh, 3500 Terrace Street, BST S 700, Pittsburgh, PA, USA. E-mail: aggarwalr@upmc.edu C The Author(s) 2021. Published by Oxford University Press on behalf of the British Society for Rheumatology. V This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com Alexander G. S. Oldroyd et al. Key words: myositis, muscle, autoantibodies, CT scanning, neoplasia, epidemiology, meta-analysis Rheumatology key messages . IIM cancer risk factors include dermatomyositis, older age, male sex, dysphagia, cutaneous ulceration and anti- TIF1c. . Cancer screening evidence is limited but CT scans may be useful. . Findings from this study can inform IIM-specific cancer screening guidelines. Introduction assimilated into a narrative review. Study selection, data extraction, quality assessment, data synthesis and ana- Idiopathic inflammatory myopathies (IIMs) are chronic lysis were all carried out in adherence to PRISMA guide- multisystem autoimmune conditions that may cause lines (for PRISMA checklist see Supplementary Material, muscle inflammation (myositis), skin manifestations and available at Rheumatology online) [8]. interstitial lung disease (ILD) [1, 2]. Adult-onset IIMs are associated with increased risk of cancer. A recent meta- Data sources analysis of such studies reported a standardized inci- A systematic literature search was carried out on dence ratio of 4.66 for DM and 1.75 for PM [3]. The gen- Medline via PubMed, Embase via OVID and Scopus. erally accepted definition of cancer associated myositis The following were used to identify appropriate studies: (CAM) is malignancy within 3 years of IIM onset. Cancer ‘myositis’, ‘neoplasm’, ‘screening’. Full-length peer remains the leading cause of death for adults with IIM reviewed articles published in English language before 8 [4–6]. January 2020 were included. Case reports, letters and Around one in four patients are diagnosed with cancer conference abstracts were excluded. References of within 3 years before or after IIM onset [4]; however, risk each identified study were also examined for further ap- varies according to the presence/absence of certain fac- propriate studies. tors [7]. Unfortunately, the data regarding risk factors are largely derived from retrospective studies with small Study selection populations, thus limiting the ability to form robust con- clusions and extrapolate to the wider population. Studies were included in the risk factor meta-analysis if Meta-analysis of existing evidence could synthesize they provided data on at least one risk factor, included results across studies and identify factors associated at least 10 IIM study subjects, and provided data on an with cancer in IIM populations, thus potentially account- IIM control group. It is important to note that risk factors ing for biases, such as publication bias and outlier stud- were assessed in comparison to each study’s wider IIM ies. Further, assimilation of evidence relating to cancer population, not the general healthy population. Eligible screening specific to CAM could provide an evidence IIM subtypes included DM, PM, anti-synthetase syn- base informing clinical screening practices and facilitate drome (ASS), immune-mediated necrotizing myopathy the formulation of cancer screening guidelines. (IMNM) and clinically amyopathic DM (CADM). Data The International Myositis Assessment and Clinical relating to inclusion body myositis were excluded due to Studies Group (IMACS), the largest international group the relationship with cancer being distinct from that of for scientific studies in myositis, began a special interest other IIM subtypes [4]. Only the study with the largest group to develop evidence-based cancer screening cohort was included where repeated studies utilized the guidelines for newly diagnosed IIM patients. This meta- same cohort data, where identifiable. analysis forms an important component in guideline for- For the review of screening practices, studies that mation. Therefore, the overall aim of this study is to assessed at least one cancer screening approach/mo- identify risk factors associated with CAM using meta- dality in an IIM population were included. analysis, and to systematically review existing evidence relating to CAM screening approaches. Data extraction Each eligible article was independently reviewed by two Methods reviewers (A.O., M.D.G., D.K., S.T., A.A., A.P. and K.K.). The title and study abstracts were reviewed to assess We performed a systematic review of factors associated eligibility/ineligibility. Preliminary full text reviews were with cancer in IIM populations and screening practices. carried out where eligibility/ineligibility could not be Evidence pertaining to factors associated with cancer decided using the title and abstract alone. Full text re- were assimilated via meta-analysis. Results of studies view of each eligible article was carried out by a single relating to cancer screening in IIM populations were reviewer. Extracted data included study type, population 2616 https://academic.oup.com/rheumatology Cancer and myositis – a meta-analysis and review studied, sample size, risk factors evaluated, number of factors with significant (<0.05) Egger’s test P-values cases (i.e. those with risk factors), controls (i.e. those [13]. without risk factors), and number of cases and controls All analysis was carried out using the statistical pro- diagnosed with cancer (excluding non-melanotic skin gramme R [14], and the meta [15] and metaphor [16] cancers). Available data (e.g. mean, S.D., median, range) packages. on continuous risk factors, such as age, in those with/ without cancer were also collected. A second reviewer Results reviewed selected studies to ensure accuracy of data extraction. The quality of studies and bias assessment A total of 7030 articles were initially identified via the lit- was carried out using the GRADE system developed by erature search and 141 were reviewed for eligibility fol- the Scottish Intercollegiate Guidelines Network, where lowing removal of ineligible papers, duplicates, case each study was given a quality assessment of either reports and reviews (Fig. 1). Sixty-seven studies were very low, low, moderate or high [9]. Studies were included in the risk factor meta-analysis (Fig. 1) and nine excluded if they were deemed to be of low or very low in the screening narrative review (two studies were quality or subject to a high risk of bias according to the included in both the meta-analysis and systematic re- GRADE system. Agreement of both reviewers was view). Table 1 displays the summary RRs and WMDs required to remove a study according to bias. The deci- calculated for each risk factor. See Supplementary sion of study inclusion/exclusion was made by a third Material for forest plots for each risk factor and further reviewer in the case of differing assessments. details of each study (Supplementary Table 1, available at Rheumatology online). Data synthesis and analysis Meta-analysis was carried out for each risk factor where IIM subtypes data from at least two eligible studies were available. DM was significantly associated with a higher risk of Investigated factors included IIM subtypes, demograph- cancer compared with other IIM subtypes [17–40]. PM ics, clinical features, laboratory parameters and autoan- [17–28, 30, 31, 33–40] and CADM [28, 35, 41] were tibodies. The denominator used in cancer risk found to be associated with significantly lower risk for estimation for each factor was the remaining IIM popula- cancer compared with remaining IIM subtypes. ASS tion of each study, not the general population. The can- subtype was a non-significant factor; however, data cer risk associated with individual ASS-related from only two eligible studies were available [21, 42]. autoantibodies (anti-Jo1, anti-PL7, anti-PL12, anti-EJ, Insufficient data were available to perform meta-analysis anti-OJ, anti-KS) was considered. Subsequently, the risk on data relating to IMNM. associated with the presence of any ASS-related auto- antibody was calculated by combining studies that com- Clinical factors including demographics and pared risk against non-ASS IIM controls. Risk ratios laboratory values (RRs) were calculated for binary variables (e.g. presence of ILD). The weighted mean difference (WMD) for each Demographics continuous variable (e.g. age) was calculated by com- Older age at time of IIM onset was found to be signifi- paring means and S.D.s. The mean (S.D.) was calculated cantly associated with increased risk of cancer [21, 22, from studies that reported only median and range using 28, 29, 32, 33, 36, 39, 42–54]. The mean age of IIM methods described by Hozo et al. [10]. onset in cancer cases was 59 years, compared with The small number of studies that reported the utility of 49 years in the non-cancer cases. Male sex was found cancer screening approaches in IIM populations pre- to be significantly associated with higher risk of cancer, cluded a meta-analysis, therefore a narrative review was compared with female sex [17, 18, 21, 22, 25–30, 32– carried out. 34, 36, 38, 39, 41–59]. Clinical risk factors Heterogeneity and study sample size analysis Sufficient data were available to quantify the cancer Heterogeneity was assessed using the standard chi- risks associated with dysphagia, cutaneous ulceration, squared test and I2 statistic. Further analysis was car- Raynaud’s phenomenon and ILD. Dysphagia, which was ried out for factors with very high levels of heterogeneity typically not objectively defined across the majority of (I2 >75%). Influence analysis (‘leave-one-out’) was car- studies, was significantly associated with higher risk of ried out to identify outlier studies, that is those with ex- cancer [22, 25, 26, 29, 32, 33, 39, 43, 46, 47, 49, 56, 59, treme effect sizes, and thus substantially contributing to 60]. Cutaneous ulceration was also significantly associ- heterogeneity. A study was considered an outlier if it ful- ated with higher risk of cancer [45, 46, 49, 50, 53]. filled the cut-off criteria proposed by Viechtbauer et al. Analysis revealed that the presence of Raynaud’s phe- [11]. nomenon was associated with a significantly lower risk Egger’s test was used to assess the influence of of cancer [22, 25, 26, 29, 30, 39, 41, 45, 46, 50, 59]. The study cohort size on calculated effect sizes [12]. ‘Trim presence of ILD, which was typically diagnosed via CT and fill’ was used to calculate adjusted effect sizes for scanning, was also associated with a significantly lower https://academic.oup.com/rheumatology 2617 Alexander G. S. Oldroyd et al. FIG. 1 PRISMA flow diagram risk of cancer [22, 26, 28–30, 32, 33, 35, 39, 43, 45, 46, non-significant factor [21, 37, 61, 62, 68, 71–76]. Large 56, 59]. proportions of the control cohorts in studies of anti- NXP2-positive cohorts were comprised of anti-TIF1c- Laboratory values positive cases. We repeated meta-analysis after remov- Lower creatine kinase (CK) [22, 27, 29, 32, 33, 38, 39, ing anti-TIF1c-positive cases. The RR of anti-NXP2 using 43, 46, 53] values were significantly associated with data from six studies with anti-TIF1c cases excluded increased cancer risk, therefore, conversely, very high was 1.47 [95% confidence interval (CI): 0.57, 3.80; I2 values were associated with lower risk of cancer. It is 0.00%], again indicating that positivity for anti-NXP2 is a important to note, however, that the mean CK level in non-significant factor for cancer relative to other auto- cancer cases (2402 IU/l) was still raised compared with antibody subtypes [21, 37, 62, 68, 71, 73]. normal values, but lower than the non-cancer group Analysis was carried out for each individual ASS- (3557 IU/l). Similarly, lower lactate dehydrogenase (LDH) related autoantibody. Anti-Jo-1 [21, 22, 25, 26, 29, 33, [22, 32, 33, 38, 39, 46, 53] values (mean LDH 766 U/l) 46, 59, 62, 71, 77] and anti-EJ [21, 62, 71, 78, 79] were were found to be associated with increased cancer risk significantly associated with reduced cancer risk. as compared with higher LDH values (mean LDH Positivity for anti-PL7 [21, 62, 71, 78, 79], anti-PL12 [21, 1078 U/l). Both alanine transaminase (ALT) [29, 38, 53] 62, 71, 78, 79], anti-OJ [21, 71, 78, 79] and anti-KS [71, and ESR [22, 29, 33, 38, 43, 45] levels were found to be 79] were non-significant factors, although limited by non-significant factors, and insufficient data were avail- small number of studies. Analysis revealed that the pres- able for aspartate aminotransferase and aldolase. ence of any ASS-related autoantibody was significantly Autoantibodies associated with lower risk of cancer [21, 22, 25, 26, 29, Anti-transcriptional intermediary factor-1 gamma (anti- 33, 59, 62, 70, 71]. TIF1c) positivity was significantly associated with Positivity for other autoantibodies, including anti-3- increased cancer risk [21, 31, 37, 41, 44, 48, 58, 60–71]. hydroxy 3-methylutaryl coenzyme A reductase (anti- Anti-nuclear matrix protein 2 (anti-NXP2) positivity was a HMGCR) [21, 62, 71, 80, 81], anti-signal recognition 2618 https://academic.oup.com/rheumatology Cancer and myositis – a meta-analysis and review TABLE 1 Meta-analysis results, including calculated effect sizes, heterogeneity and publication bias for each factor Domain Factor RR/WMD (95% CI) GRADE Heterogeneity Egger’s certainty test, ratinga P-value P-value I2, % Subtypeb DM 2.21 (1.78, 2.77) Moderate 0.134 24.8 0.162 PM 0.49 (0.37, 0.65) Moderate 0.061 34.0 0.603 CADM 0.44 (0.20, 0.97) Low 0.751 0.0 — ASS 0.28 (0.00, 6554.79) Low 0.017 82.4 — Demographics Age 11.19 (9.29, 13.08) High 0.001 56.1 0.859 Male sex 1.53 (1.34, 1.75) High 0.101 24.2 0.081 Clinical features Dysphagia 2.09 (1.21, 3.60) High <0.0001 83.6 0.310 Cutaneous ulceration 2.73 (1.33, 5.59) Moderate 0.389 3.1 — Raynaud’s phenomenon 0.61 (0.39, 0.95) Moderate 0.698 0.0 0.055 ILD 0.49 (0.32, 0.76) High 0.011 52.3 0.381 Blood parameters CK 1189.96 ( 2132.74, 247.18) Moderate <0.0001 84.5 0.182 LDH 336.52 ( 514.40, 158.64) Moderate 0.093 44.8 — ALT 36.29 ( 313.18, 385.77) Low 0.001 85.5 — ESR 4.4 ( 6.2, 14.9) Low 0.028 60.2 — Autoantibodies Anti-TIF1c 4.68 (3.37, 6.48) High <0.0001 68.8 0.543 Anti-NXP2 1.16 (0.73, 1.87) Moderate 0.278 17.4 0.271 Anti-c 1.59 (0.33, 7.74) Low 0.217 34.5 — Anti-HMGCR 0.55 (0.19, 1.61) Low 0.277 21.6 — Anti-MDA5 0.17 (0.02, 1.28) Low 0.457 0.0 — Anti-Mi2 1.05 (0.28, 3.92) Low 0.381 5.6 — Anti-SRP 0.40 (0.14, 1.21) Low 0.790 0.0 — Any ASS antibody 0.41 (0.26, 0.64) High 0.746 0.0 <0.001 Anti-Jo1 0.45 (0.25, 0.84) High 0.700 0.0 0.051 Anti-PL7 0.68 (0.15, 3.07) Low 0.541 0.0 — Anti-PL12 1.59 (0.89, 2.86) Low 0.789 0.0 — Anti-EJ 0.17 (0.07, 0.44) Low 0.964 0.0 — Anti-OJ 1.56 (0.68, 3.52) Low 0.870 0.0 — Anti-KS 1.23 (0.05, 30.12) Very low 0.717 0.0 — MSA negative 0.89 (0.50, 1.59) Low 0.073 50.4 — ANA positivity 0.91 (0.58, 1.41) Low <0.0001 75.0 0.113 a Grading of Recommendations, Assessment, Development and Evaluations certainty rating: very low (the true effect is probably markedly different from the estimated effect), low (the true effect might be markedly different from the estimated effect), moderate (the authors believe that the true effect is probably close to the estimated effect), and high (the authors have a lot of confidence that the true effect is similar to the estimated effect). bThe risk of cancer for each IIM subtype is estimated against each study’s wider IIM population, not the general population. ALT: alanine transaminase; ASS: anti-syn- thetase syndrome; CADM: clinically amyopathic dermatomyositis; CI: confidence interval; CK: creatine kinase; DM: derm- atomyositis; HMGCR: 3-hydroxy 3-methylutaryl coenzyme A reductase; ILD: interstitial lung disease; LDH: lactate dehydrogenase; MDA5: melanoma differentiation-associated gene 5; MSA: myositis specific autoantibody; NXP2: nuclear matrix protein 2; RR: risk ratio; SAE1: small ubiquitin-like modifier-1 activating enzyme; SRP: signal recognition particle; TIF1c: transcriptional intermediary factor-1 gamma; WMD: weighted mean difference. particle (anti-SRP) [62, 71, 80], anti-small ubiquitin-like study each fulfilled the outlier criteria for CK [53], ESR modifier-1 activating enzyme (anti-SAE1) [21, 62, 71], [45] and ALT [53]. Adjusted WMD after removal of data anti-melanoma differentiation-associated gene 5 (anti- from outlier publication was calculated and did not MDA5) [21, 44, 48, 62, 71] or anti-Mi2 [21, 41, 58, 62, change overall relationships (Supplementary Table 2, 64, 71], were identified as non-significant factors for available at Rheumatology online). cancer. Both myositis specific autoantibody (MSA) nega- Significant publication bias was observed with ‘any tivity [21, 30, 45, 62, 71] and ANA positivity [22, 26, 28, ASS-antibody’. Adjusted RR following ‘trim and fill’ ana- 30, 39, 41, 43, 46, 50, 53, 54, 56, 82] were non-signifi- lysis with six added studies was 0.46 (95% CI: 0.23, cant factors. 0.93). Heterogeneity and publication bias Cancer screening utility review Table 1 displays the standard chi-squared test results Nine studies [40, 45, 83–89] relating to utility of cancer and I2 statistic for heterogeneity of each analysed factor. screening approaches in IIM populations were identified. Influence analysis aimed to identify outlier studies for Table 2 displays the details of each study. A total of 90 risk factors with very high (>75%) heterogeneity. One cancers were identified via screening across 1033 https://academic.oup.com/rheumatology 2619 TABLE 2 Details of identified studies reporting utility of cancer screening investigations in IIM populations 2620 Study Country Study type IIM subtypes Population Screening mo- Timing of No. cancer Control modality No. cancer Comments included size dality assessed screening cases identified cases identified Kidambi et al. USA Retrospective DM 79 in total Upper and lower Mean 6.8 years 0 NA NA [83] Upper GI GI endoscopy (S.D. 6.6) after 2 cases of endoscopy DM onset Barrett’s 47 oesophagus Lower GI 10 adenoma endoscopy Alexander G. S. Oldroyd et al. 67 18 18 Maliha et al. [84] Canada Retrospective DM 31, PM 1, over- 63 F-FDG PET/CT ‘Average’ time of 0 ‘Conventional’, 3 F-FDG PET/CT lap 25, IBM 1, 9 months after physical and breast cancer via lead to more orbital 1, unspeci- IIM diagnosis gynaecological mammog- biopsies com- fied subtype 4 examination, raphy, squa- pared with con- CBC, serum mous cell ventional biochemistry, carcinoma via screening (8 vs LFTs, serum examination, 5) protein electro- multiple mye- phoresis, urin- loma via blood alysis, CXR, tests gastroscopy, colonoscopy, CT of thorax, abdomen and pelvis, mam- mography, endovaginal USS, serum tu- mour markers Leatham et al. USA Retrospective DM 400 Median 4.2 years 29 cancers NA NA 2 cancers (breast [85] (IQR 1.7–8.0.6) across 27 cancer and between symp- patients diag- DLBCL) were tom onset and nosed after DM diagnosed via screening onset repeat ‘blind 17 cancers (16 screening’. patients) diag- Increasing age nosed by blind was only iden- screening tified cancer CT abdomen 4 risk factor. Mammography 3 CBC 3 Colonoscopy 2 PSA 2 CT thorax 1 CT pelvis 1 Huang et al. [86] China Retrospective DM and PM 129 WBMRI Mean disease 5, all DM NA NA PM 30, DM 99 duration 3 NPC 30.8 months 1 ovarian (S.D. 47.9, range 1 thyroid cancer (continued) https://academic.oup.com/rheumatology TABLE 2 Continued Study Country Study type IIM subtypes Population Screening mo- Timing of No. cancer Control modality No. cancer Comments included size dality assessed screening cases identified cases identified 10 days to 19 years) Whitmore et al. USA Retrospective DM 14 Serum CA-125 Median disease 4 ovarian cancer NA NA Pre-diagnosis [87] duration serum CA-125 15.5 months levels were (range 7–24) found to be higher in 2 https://academic.oup.com/rheumatology (50%) of ovar- ian cancer cases. Normal serum CA-125 levels were seen in all controls. Sparsa et al. France Retrospective DM and PM 40 History, physical Not reported for Total 122 investi- NA NA CT TAP revealed [45] and pelvic whole cohort gations, 30 most ‘blind’ examination, Screening revealed screening can- CBC, ESR, occurred be- malignancy cers - 5/18 general chem- tween 35 tests were (28%) were istry screen, 12 months ‘directed’, 19 positive LFTs, CXR, prior to and (54%) were mammography, 8 months after positive CT TAP, upper IIM onset in 87 tests were and lower GI cancer cases ‘blind’, 11 endoscopy, (13%) were ‘small bowel positive radiologic examination’, thyroid imag- ing, MRI, PET- CT, cancer- associated anti- gens, bone marrow biopsy, laparotomy 18 18 Selva- Spain Prospective 49 DM, 55 F-FDG PET/CT Within 6 months Positive in 7 CT abdomen and Positive in 9 F-FDG PET/CT O’Callaghan 6 PM period after IIM cases (1 false- pelvis, mam- cases (2 false- PPV was 86%, et al. [88] diagnosis positive), nega- mography, positive) NPV was 94% tive in 44 cases gynaecologic 5 breast, 1 lung, 1 (3 false-nega- examination, pancreas, 1 tives) and in- ovarian USS, vagina, 1 conclusive in 4 tumour markers colon cases (CA-125, CA- Negative in 46 19.9, CEA, cases (2 false- PSA) negatives) Amoura et al. France Retrospective 50 DM, 102 CEA > 5 ng/ml Not reported CEA increased in NA NA [89] 52 PM CA15-3 > 25 4 patients, no units/ml cancer diagnoses (continued) Cancer and myositis – a meta-analysis and review 2621 TABLE 2 Continued 2622 Study Country Study type IIM subtypes Population Screening mo- Timing of No. cancer Control modality No. cancer Comments included size dality assessed screening cases identified cases identified CA19-9 > 37 CA15-3 increased units/ml in 22, 2 cancer CA125 > 35 units/ diagnoses ml CA19-9 increased in 11 patients, 3 cancer diagnoses Alexander G. S. Oldroyd et al. CA125 increased in 8 patients, 5 cancer diagnoses CA19-9 and CA125 were both increased in 3 patients, all 3 were diagnosed with cancer Lim et al. [40] Taiwan Retrospective 98 DM, 152 CEA > 5 ng/ml Mean 6.1 years CA15-3 increased 8 (89%) of the 9 53 PM CA125 > 35 units/ (S.D. 5.7) after in 9 patients, no with elevated ml IIM onset cancer CA15-3 levels CA19-9 > 34 diagnoses developed ILD units/ml CA125 increased CA15-3 > 25 in 18 patients, units/ml 1 cancer AFP >12 ng/ml diagnosis CA19-9 increased in 10, 1 cancer diagnosis AFP increased in 4, no cancer diagnoses CEA increased in 8, 3 cancer diagnoses 18 F-FDG PET/CT: 18F-fluorodeoxyglucose PET/CT; AFP: alpha fetoprotein; CA: carbohydrate antigen; CBC: complete blood count; CEA: carcinoembryonic antigen; CT TAP: com- puted tomography thorax, abdomen and pelvis; CXR: chest X-ray radiograph; DLBCL: diffuse large B cell lymphoma; GI: gastrointestinal; IIM: idiopathic inflammatory myopathy; ILD: interstitial lung disease; IQR: interquartile range; LFT: liver function tests; NA: not applicable; NPC: nasopharyngeal carcinoma; NPV: negative predictive value; PPV: positive predictive value; PSA: prostate-specific antigen; USS: ultrasound scan; WBMRI: whole body magnetic resonance imaging. https://academic.oup.com/rheumatology Cancer and myositis – a meta-analysis and review patients. Studies were carried out across a number of were significantly associated with cancer in the study by countries, including the USA, Canada, Taiwan, China, Amoura et al.; 11 cases had raised levels and three sub- France and Spain, and widely ranging intervals between sequently developed cancer (OR 4.5; 95% CI: 1.00, IIM onset/diagnosis and screening were reported. All but 18.7; P-value 0.018). Raised CA19-9 levels were not one study was retrospective. Study population sizes found to be associated with cancer in the study by Lim ranged between 14 and 400. A wide variety of cancers et al. however. Of note, Amoura et al. reported that were diagnosed, including but not limited to breast can- three cases had raised levels of both CA19-9 and cer, squamous cell carcinoma, multiple myeloma, ovar- CA125 and all of these were subsequently diagnosed ian cancer, lymphoma, lung cancer and oesophageal with cancer (OR 86.3; 95% CI: 4, 1832; P-value cancer. <0.0001). Lim et al. also reported no association be- The utility of ‘blind screening’ (i.e. investigations car- tween raised AFP levels and cancer. Interestingly, Lim ried out in the absence of target symptoms) was et al. reported an association between CA15-3 levels reported by Leatham et al. [85] and Sparsa et al. [45]. and the development of ILD; eight (89%) of the nine Leatham et al. identified 17 out of 48 cancer patients patients with increased CA15-3 levels were diagnosed diagnosed with cancer via blind screening modalities with ILD. after DM onset. CT scanning of the thorax, abdomen or pelvis detected the most cancer diagnoses (6/17, 38%), Discussion followed by mammography (3/17, 18%). Sparsa et al. reported the identification of 30 cancers via 122 investi- This meta-analysis has quantified the relationship be- gations. Thirty-five investigations were ‘directed’ (i.e. ini- tween 30 clinical factors and the risk of cancer in IIM tiated due to the presence of target symptoms) and patients. Fifteen factors significantly associated with resulted in the identification of 19 (54%) cancers. In con- cancer risk were identified. Existing evidence relating to trast, 87 investigations were blind and identified 11 the utility of cancer screening in IIM populations was (13%) cancers. Again, CT scanning of the thorax, abdo- also reviewed, providing information useful for the future men and pelvis was the single investigation that formation of cancer screening guidelines. detected the most cancers (5/18, 28%). DM, increasing age, male sex, dysphagia, cutaneous The utility of 18F-FDG PET/CT was reported by Maliha ulceration and the presence of anti-TIF1c were all asso- et al. [84] and Selva-O’Callaghan et al. [88]. Maliha et al. ciated with increased cancer risk. The magnitude of risk reported that fluorodeoxyglucose (18F-FDG) PET/CT of cancer was greatest for those positive for anti-TIF1c, scans revealed no further cancer diagnoses and actually with a fourfold increased risk. Very high LDH or CK val- lead to more biopsies, compared with ‘conventional’ ues were associated with reduced cancer risk. screening (see Table 2 for details). Similarly, Selva- PM and CADM subtypes were associated with lower O’Callaghan reported that single 18F-FDG PET/CT scans risk of cancer compared with other subtypes. However, were comparable to a large number of conventional the risk of cancer in PM and CADM cases may be screening investigations, which included complete phys- reduced, but the risk is still raised compared with the ical examination, laboratory tests (complete blood count general population, as previously identified [3]. and serum chemistry panel), thoraco-abdominal CT ASS subtype was a non-significant factor for cancer; scan, tumour markers [carbohydrate antigen-125 however, this was based on data from only two studies. (CA125), CA19-9, carcinoembryonic antigen (CEA), pros- The presence of ILD or any ASS-related antibody, in tate-specific antigen], gynaecological examination, ovar- particular anti-Jo1 and anti-EJ, were significantly associ- ian ultrasonography and mammography. ated with lower cancer risk. ASS is characterized by ILD The screening utility of CA125 was demonstrated by and the presence of any ASS-related antibody, therefore Amoura et al. [89] and Whitmore et al. [87]. Amoura it may be concluded that ASS patients are at significant- et al. demonstrated that increased levels were signifi- ly lower cancer risk compared with other IIM subtypes. cantly associated with subsequent cancer diagnoses Insufficient evidence was available to include IMNM (OR 29.7; 95% CI: 8.2, 106.6; P-value <0.0001). subtype in the meta-analysis. However, meta-analysis Whitmore et al. also demonstrated the utility of normal was possible for anti-SRP and anti-HMGCR, both values—no study participant with normal CA125 levels IMNM-specific autoantibodies. Positivity for either anti- was subsequently diagnosed with cancer during the SRP or anti-HMGCR were non-significant factors for study period. In contrast, Lim et al. concluded that cancer. Additionally, very high CK levels, which are also CA125 testing was not useful for detection of cancer typically observed in IMNM cases, were associated with [40]. Eighteen participants had raised CA125 levels and reduced cancer risk. A small number of studies have only one (6%) was subsequently diagnosed with cancer. reported increased risk of cancer in IMNM patients Additionally, 53 participants had normal CA125 levels compared with the general population; however, the risk and two (4%) were diagnosed with cancer. may be dependent on autoantibody status, as reported Both Amoura et al. [89] and Lim et al. [40] reported by Allenbach et al. [80], where anti-HMGCR positivity the screening utility of CEA, CA15-3 and CA19-9 was associated with increased cancer risk and anti-SRP (Table 2). Raised CEA or CA15-3 levels were not associ- positivity was not. An increased cancer risk associated ated with cancer in each study. Raised CA19-9 levels with anti-HMGCR positivity compared with the general https://academic.oup.com/rheumatology 2623 Alexander G. S. Oldroyd et al. population was, however, not found by Tiniakou et al. required to fully delineate the role of 18F-FDG PET/CT [90]. Overall, the relationship between IMNM and cancer scanning as a screening strategy for cancer in IIM remains unclear, and further research in larger cohorts is patients. warranted. As previously mentioned, all results and findings in Anti-NXP2 positivity was not associated with cancer this study pertain only in comparison to IIM patients, not in this meta-analysis even after removal of anti-TIF1c- the general population. Future research and meta-ana- positive cases, where possible. Previous studies have, lysis may consider delineating the cancer risk of appro- however, highlighted the increased risk of anti-NXP2 priate factors in comparison to the general population. positivity compared with the general population, for ex- One major potential limitation to this study is the vary- ample Yang et al. reported a cancer risk standardized ing MSA detection methods employed by different stud- incidence ratio of 8.14 compared with the general popu- ies. This introduces the risk of varying accuracy of MSA lation [21]. It is perhaps, therefore, still appropriate to detection, thus affecting the calculated effect sizes. consider anti-NXP2 positivity a cancer risk factor when Further, substantial heterogeneity potentially limits the considering comparison to the general population. clinical translation of variables studied. Publication bias Further research to fully delineate the cancer risk asso- was observed with any ASS-related antibody, thus high- ciated with anti-NXP2 positivity is warranted. lighting potential inaccuracy of calculated effect sizes. Few previous studies have investigated the utility of Recent advances in understanding raise the possibility cancer screening approaches in IIM populations; how- that PM cases may actually represent other subtypes, ever, a number of conclusions can be drawn. such as IMNM or other neuromuscular disorders [94– Firstly, imaging of internal organs via CT scanning of 96], thus potentially limiting the accuracy of the esti- the thorax, abdomen and pelvis appeared to yield a mated cancer risk associated with PM. Calculation of high proportion of cancers. CT scanning is a readily the cancer risk associated with connective tissue dis- available low-cost investigation and therefore represents ease-associated IIM (overlap IIM) was not possible due a potentially useful method of screening. to varying classification. A number of potential risk fac- Secondly, CA125 levels may potentially be useful in tors such as ethnicity, arthralgia, arthritis and fever were stratifying patients’ ovarian cancer risk. It is important to not included in this meta-analysis due to unavailability of note, however, that the evidence is overall weak, with objective data. No studies addressed whether or not only three studies reporting relevant results. repeated cancer screening is beneficial in identifying Thirdly, neither of the two included studies demon- cancer; evidence on this important topic will impact strated that 18F-FDG PET/CT scanning leads to a higher screening practices, especially in patients where no can- yield of cancer diagnosis [84, 88]. The study by Selva- cer was diagnosed via initial screening. The potential O’Callaghan et al., however, indicated that 18F-FDG interaction of the presence of multiple risk factors and PET/CT scanning was comparable to a wide panel of their impact upon stratification of cancer risk in IIM has extensive screening investigations in ability to detect never been evaluated. The small number of studies that cancers. This indicates that a single 18F-FDG PET/CT report the utility of cancer screening investigations high- scan may potentially negate the need for numerous lights the need for further research in this area. investigations. It is important to note the small popula- tion sizes in the studies by Maliha et al. [84] and Selva- O’Callaghan et al. [88] and non-stratification according Conclusion to the presence of risk factors, thus precluding extrapo- This meta-analysis has quantified the risk of cancer lation of utility of 18F-FDG PET/CT in IIM patients with associated with a large number of clinical risk factors risk factors. The higher number of biopsies performed and MSAs, which can inform cancer screening practices following 18F-FDG PET/CT without subsequent cancer for IIM patients. In addition, the systematic review of diagnoses, as reported by Maliha et al., is also a poten- available evidence related to utility of cancer screening tial disadvantage. 18F-FDG PET/CT can provide poten- investigations, although limited, can also inform clinical tially useful IIM-specific clinical information relating to decisions and aid guideline development in this area. ILD and myositis [91]. Further, a single 18F-FDG PET/CT Overall, these results can inform the development of scan can result in lower out of pocket expenses for cancer screening guidelines, thus potentially leading to patients (US $127 less), compared with a broad panel of earlier cancer diagnosis and improved patient screening investigations (i.e. CT, tumour markers, faecal outcomes. occult blood, mammography, ovarian ultrasonography) [92]. However, a small but potentially important radiation Acknowledgements exposure of 25 mSv is associated with an 18F-FDG PET/ CT scan, compared with 14–19 mSv with a standard This study was developed and conducted under the whole-body CT scan [93]. 18F-FDG PET/CT may there- International Myositis Assessment and Clinical Studies fore represent a cost-effective single investigation that Group (IMACS). R.A. conceived and designed the study. can identify underlying malignancy and detect ILD and A.O., M.D.G., D.K., S.T., A.A., A.P. and K.K. carried out myositis, thus removing the need for further multiple the literature review and data extraction. A.O. carried screening investigations. Further evidence is, however, out all data analysis. A.O. led manuscript preparation. 2624 https://academic.oup.com/rheumatology Cancer and myositis – a meta-analysis and review All authors reviewed results, critically appraised the 2 Chinoy H, Cooper RG. Polymyositis and manuscript and approved the final version. dermatomyositis. In: Chinoy H and Cooper RG, eds. Oxford Textbook of Rheumatology. Oxford: Oxford Funding: This report includes independent research sup- University Press, 2013:1009–20. ported by the National Institute for Health Research 3 Qiang JK, Kim WB, Baibergenova A, Alhusayen R. Risk Biomedical Research Centre Funding Scheme. The of malignancy in dermatomyositis and polymyositis: a views expressed in this publication are those of the systematic review and meta-analysis. J Cutan Med Surg authors and not necessarily those of the National Health 2017;21:131–6. Service, the National Institute for Health Research or the 4 Dobloug GC, Garen T, Brunborg C, Gran JT, Molberg Ø. Department of Health. Survival and cancer risk in an unselected and complete Norwegian idiopathic inflammatory myopathy cohort. This work was supported by funding from Myositis UK Semin Arthritis Rheum 2015;45:301–8. (to A.O); Versus Arthritis (Award Number: 21993 to A.O.); the National Institute for Health Research Manchester 5 Nun~o-Nun~o L, Joven BE, Carreira PE et al. Mortality and Biomedical Research Centre (to A.O. and H.C.); the prognostic factors in idiopathic inflammatory myositis: a retrospective analysis of a large multicenter cohort of National Institute for Health Research University College Spain. Rheumatol Int 2017;37:1853–61. London Hospitals Biomedical Research Centre (to P.M.M.) and Bath Institute of Rheumatic Diseases (to 6 Dobloug GC, Svensson J, Lundberg IE, Holmqvist M. Mortality in idiopathic inflammatory myopathy: results S.T.). from a Swedish nationwide population-based cohort Disclosure statement: R.A. served as a consultant for study. Ann Rheum Dis 2018;77:40–7. Kezar, Csl Behring, AstraZeneca, Octapharma, BMS, 7 Moghadam-Kia S, Oddis CV, Ascherman DP, Aggarwal Pfizer, Janssen, Mallinckrodt, Alexion, Q32, Argenx, R. Risk factors and cancer screening in myositis. Rheum Boehringer-Ingelheim, Corbus and EMD-Serono; and Dis Clin North Am 2020:46;565–76. received research funding from Pfizer, BMS, Genentech, 8 Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Kezar, Csl Behring and Mallinckrodt. L.C. served on an Group. Preferred reporting items for systematic reviews advisory board and received grant funding from and meta-analyses: the PRISMA statement. PLoS Med Boerhinger Ingelheim; served on an advisory board for 2009;6:e1000097. Eicos Sciences; served on an advisory board for Bristol- 9 Scottish Intercollegiate Guidelines Network. SIGN 50. A Myers Squibb; received consulting fees for Mitsubishi guideline developer’s handbook, 2011. http://www.sign. Tanabe; and serves on a data safety monitoring board ac.uk/assets/sign50_2011.pdf (22 January 2021, date for Reata. M.D.G. received a research grant from last accessed). Bristol-Myers Squibb for unrelated work; and consulting 10 Hozo SP, Djulbegovic B, Hozo I. Estimating the mean fees from Dysimmune Diseases Foundation. P.M.M. and variance from the median, range, and the size of a received consulting/speaker’s fees from Abbvie, BMS, sample. BMC Med Res Methodol 2005;5:2020. Celgene, Eli Lilly, Janssen, MSD, Novartis, Orphazyme, 11 Viechtbauer W, Cheung MW-L. Outlier and influence Pfizer, Roche and UCB, outside the submitted work. diagnostics for meta-analysis. Res Synth Methods 2010; R.A.V. received a research grant from Pfizer. H.C. 1:112–25. received research grants, travel grants, consultancy or 12 Egger M, Smith GD, Schneider M, Minder C. Bias in speaker honoraria from AbbVie, Amgen, BMS, Biogen, meta-analysis detected by a simple, graphical test. Br Janssen, Lilly, Novartis and UCB. V.P.W. served as a Med J 1997;315:629–34. consultant for Kezar, CSL Behring, AstraZeneca, 13 Duval S, Tweedie R. Trim and fill: a simple funnel-plot- Octapharma, Pfizer, Janssen, Neovacs and Idera; and based method of testing and adjusting for publication received research funding from Pfizer, CSL Behring and bias in meta-analysis. Biometrics 2000;56:455–63. Corbus. 14 R Core Team. R: A language and environment for statistical computing, 2014. Available from: http://www.r- Data availability statement project.org/ (22 January 2021, date last accessed). 15 Balduzzi SRücker, G Schwarzer, G. How to perform a There are no new data associated with this article. meta-analysis with R: a practical tutorial. Evid Based Ment Health 2019;36:1–48. Supplementary data 16 Viechtbauer W. Conducting meta-analysis in R with the metafor package. J Stat Softw 2010;36:1–48. Supplementary data are available at Rheumatology online. 17 Huang YL, Chen YJ, Lin MW et al. Malignancies associated with dermatomyositis and polymyositis in Taiwan: a nationwide population-based study. Br J References Dermatol 2009;161:854–60. 1 Oldroyd A, Lilleker J, Chinoy H. Idiopathic inflammatory 18 Chen Y-J, Wu C-Y, Huang Y-L et al. Cancer risks of myopathies - a guide to subtypes, diagnostic approach dermatomyositis and polymyositis: a nationwide cohort and treatment. Clin Med (Northfield IL) 2017;17:322–8. study in Taiwan. Arthritis Res Ther 2010;12:R70. https://academic.oup.com/rheumatology 2625 Alexander G. S. Oldroyd et al. 19 Kuo CF, See LC, Yu KH, Chou IJ et al. Incidence, malignancy in idiopathic inflammatory myopathy. Korean cancer risk and mortality of dermatomyositis and J Intern Med 2014;29:814–21. polymyositis in Taiwan: a nationwide population study. 34 Wakata N, Kurihara T, Saito E, Kinoshita M. Br J Dermatol 2011;165:1273–9. Polymyositis and dermatomyositis associated with 20 Hill CL, Zhang Y, Sigurgeirsson B et al. Frequency of malignancy: a 30-year retrospective study. Int J specific cancer types in dermatomyositis and Dermatol 2002;41:729–34. polymyositis: a population-based study. Lancet 2001; 35 Ikeda S, Arita M, Misaki K et al. Incidence and impact of 357:96–100. interstitial lung disease and malignancy in patients with 21 Yang H, Peng Q, Yin L et al. Identification of multiple polymyositis, dermatomyositis, and clinically amyopathic cancer-associated myositis-specific autoantibodies in dermatomyositis: a retrospective cohort study. idiopathic inflammatory myopathies: a large longitudinal Springerplus 2015;4:240. cohort study. Arthritis Res Ther 2017;19: 1–9. 36 Maoz CR, Langevitz P, Livneh A et al. High incidence of 22 So MW, Koo BS, Kim YG, Lee CK, Yoo B. Idiopathic malignancies in patients with dermatomyositis and inflammatory myopathy associated with malignancy: a polymyositis: an 11-year analysis. Semin Arthritis Rheum retrospective cohort of 151 Korean patients with 1998;27:319–24. dermatomyositis and polymyositis. J Rheumatol 2011;38: 37 Ceribelli A, Isailovic N, De Santis M et al. Myositis- 2432–5. Nov specific autoantibodies and their association with 23 Chen YJ, Wu CY, Shen JL. Predicting factors of malignancy in Italian patients with polymyositis and malignancy in dermatomyositis and polymyositis: a case- dermatomyositis. Clin Rheumatol 2017;36:469–75. control study. Br J Dermatol 2001;144:825–31. 38 Lee SW, Jung SY, Park MC, Park YB, Lee SK. 24 Ishizuka M, Watanabe R, Ishii T et al. Long-term follow- Malignancies in Korean patients with inflammatory up of 124 patients with polymyositis and myopathy. Yonsei Med J 2006;47:519–23. dermatomyositis: statistical analysis of prognostic factors. Mod Rheumatol 2016;26:115–20. 39 András C, Bodoki L, Nagy-Vincze M et al. Retrospective analysis of cancer-associated myositis patients over the 25 Neri R, Simone B, Iacopetti V et al. Cancer-associated past 3 decades in a Hungarian myositis cohort. Pathol myositis: a 35-year retrospective study of a monocentric Oncol Res 2020;26:1749–55. cohort. Rheumatol Int 2014;34:565–9. 40 Lim CH, Tseng C-W, Lin C-T et al. The clinical 26 Azuma K, Yamada H, Ohkubo M et al. Incidence and application of tumor markers in the screening of predictive factors for malignancies in 136 Japanese malignancies and interstitial lung disease of patients with dermatomyositis, polymyositis and clinically dermatomyositis/polymyositis patients: a retrospective amyopathic dermatomyositis. Mod Rheumatol 2011;21: study. SAGE Open Med 2018;6:205031211878189–8. 178–83. 41 Ikeda N, Takahashi K, Yamaguchi Y et al. Analysis of 27 Meena AK, Sundaram C, Uppin M et al. Incidence of dermatomyositis-specific autoantibodies and clinical malignancies in biopsy-proven inflammatory myopathy. characteristics in Japanese patients. J Dermatol 2011; Neurol India 2013;61:152–5. 38:973–9. 28 Antiochos BB, Brown LA, Li Z, Tosteson TD et al. 42 Lilleker JB, Vencovsky J, Wang G et al. The Malignancy is associated with dermatomyositis but not EuroMyositis registry: an international collaborative tool polymyositis in northern New England, USA. J to facilitate myositis research. Ann Rheum Dis 2018;77: Rheumatol 2009;36:2704–10. 30–9. 29 Fang YF, Wu YJJ, Kuo CF, Luo SF, Yu KH. Malignancy 43 Hida A, Yamashita T, Hosono Y et al. Anti-TIF1-c in dermatomyositis and polymyositis: analysis of 192 antibody and cancer-associated myositis. Neurology patients. Clin Rheumatol 2016;35:1977–84. 2016;87:299–308. 30 Limaye V, Luke C, Tucker G et al. The incidence and 44 Ogawa-Momohara M, Muro Y, Mitsuma T et al. Strong associations of malignancy in a large cohort of patients correlation between cancer progression and anti-transcription with biopsy-determined idiopathic inflammatory myositis. intermediary factor 1c antibodies in dermatomyositis patients. Rheumatol Int 2013;33:965–71. Clin Exp Rheumatol 2018;36:990–5. 31 Kang EH, Nakashima R, Mimori T et al. Myositis 45 Sparsa A, Liozon E, Herrmann F et al. Routine vs autoantibodies in Korean patients with inflammatory extensive malignancy search for adult dermatomyositis myositis: anti-140-kDa polypeptide antibody is primarily and polymyositis: a study of 40 patients. Arch Dermatol associated with rapidly progressive interstitial lung 2002;138:885–90. disease independent of clinically amyopathic dermatomyositis. BMC Musculoskelet Disord 2010;11: 46 Ponyi A, Constantin T, Garami M et al. Cancer- 223. associated myositis: clinical features and prognostic signs. Ann N Y Acad Sci 2005;1051:64–71. 32 Kang EH, Lee SJ, Ascherman DP et al. Temporal relationship between cancer and myositis identifies two 47 Bohan A, Peter JB, Bowman RL, Pearson CM. distinctive subgroups of cancers: impact on cancer risk Computer-assisted analysis of 153 patients with and survival in patients with myositis. Rheumatology polymyositis and dermatomyositis. Medicine (Oxford) 2016;55:1631–41. (Baltimore)1977;56:255–86. 33 Kim NR, Nam EJ, Kang JW et al. Complex repetitive 48 Hoshino K, Muro Y, Sugiura K et al. Anti-MDA5 and discharge on electromyography as a risk factor for anti-TIF1-c antibodies have clinical significance for 2626 https://academic.oup.com/rheumatology Cancer and myositis – a meta-analysis and review patients with dermatomyositis. Rheumatology 2010;49: 63 Fiorentino DF, Kuo K, Chung L et al. Distinctive 1726–33. cutaneous and systemic features associated with 49 Souza FHC D, Shinjo SK. Newly diagnosed antitranscriptional intermediary factor-1c antibodies in dermatomyositis in the elderly as predictor of adults with dermatomyositis. J Am Acad Dermatol 2015; malignancy. Rev Bras Reumatol 2012;52:713–21. 72:449–55. 50 Galimberti F, Li Y, Fernandez AP. Clinically amyopathic 64 Fujimoto M, Murakami A, Kurei S et al. Enzyme-linked dermatomyositis: clinical features, response to immunosorbent assays for detection of anti- medications and malignancy-associated risk factors in a transcriptional intermediary factor-1 gamma and anti-Mi- specific tertiary-care-centre cohort. Br J Dermatol 2016; 2 autoantibodies in dermatomyositis. J Dermatol Sci 174:158–64. 2016;84:272–81. 51 Hunger RE, Dürr C, Brand CU. Cutaneous 65 Labrador-Horrillo M, Martı́nez MA, Selva-O’Callaghan A leukocytoclastic vasculitis in dermatomyositis suggests et al. Anti-TIF1c antibodies (anti-p155) in adult patients malignancy. Dermatology 2001;202:123–6. with dermatomyositis: comparison of different diagnostic assays. Ann Rheum Dis 2012;71:993–6. 52 El-Azhary RA, Pakzad SY. Amyopathic dermatomyositis: retrospective review of 37 cases. J Am Acad Dermatol 66 Chinoy H, Fertig N, Oddis CV, Ollier WER, Cooper RG. 2002;46:560–5. The diagnostic utility of myositis autoantibody testing for predicting the risk of cancer-associated myositis. Ann 53 Fardet L, Dupuy A, Gain M et al. Factors associated with Rheum Dis 2007;66:1345–9. underlying malignancy in a retrospective cohort of 121 patients with dermatomyositis. Medicine (Baltimore) 67 Trallero-Araguás E, Labrador-Horrillo M, Selva- 2009;88:91–7. O’Callaghan A et al. Cancer-associated myositis and anti-p155 autoantibody in a series of 85 patients with 54 Liu WC, Ho M, Koh W-P et al. An 11-year review of idiopathic inflammatory myopathy. Medicine (Baltimore) dermatomyositis in Asian patients. Ann Acad Med 2010;89:47–52. Singapore 2010;39:843–7. 68 Best M, Jachiet M, Molinari N, the Study Group of 55 Okayasu I, Mizutaniy H, Kurihara H, Yanagisawa F. Systemic Diseases in Dermatology (EMSED): Étude des Cancer in collagen disease. A statistical analysis by Maladies Systémiques en Dermatologie et al. Distinctive reviewing the annual of pathological autopsy cases cutaneous and systemic features associated with (Nippon Boken Shuho) in Japan. Cancer 1984;54: specific antimyositis antibodies in adults with 1841–4. dermatomyositis: a prospective multicentric study of 117 56 Koh ET, Seow A, Ong B et al. Adult onset polymyositis/ patients. J Eur Acad Dermatology Venereol 2018;32: dermatomyositis: clinical and laboratory features and 1164–72. treatment response in 75 patients. Ann Rheum Dis 1993; 69 Kaji K, Fujimoto M, Hasegawa M et al. Identification of a 52:857–61. novel autoantibody reactive with 155 and 140 kDa 57 Aussy A, Fréret M, Gallay L, the OncoMyositis Study nuclear proteins in patients with dermatomyositis: an Group et al. The IgG2 isotype of anti-transcription association with malignancy. Rheumatology 2007;46: intermediary factor 1c autoantibodies is a biomarker of 25–8. cancer and mortality in adult dermatomyositis. Arthritis 70 Fujikawa K, Kawakami A, Kaji K et al. Association of Rheumatol 2019;71:1360–70. distinct clinical subsets with myositis-specific 58 Hamaguchi Y, Kuwana M, Hoshino K et al. Clinical autoantibodies towards anti-155/140-kDa polypeptides, correlations with dermatomyositis-specific anti-140-kDa polypeptides, and anti-aminoacyl tRNA autoantibodies in adult Japanese patients with synthetases in Japanese patients with dermatomyositis: dermatomyositis: a multicenter cross-sectional study. a single-centre, cross-sectiona. Scand J Rheumatol Arch Dermatol 2011;147:391–8. 2009;38:263–7. 59 Liu Y, Xu L, Wu H et al. Characteristics and predictors 71 Betteridge Z, Tansley S, Shaddick G et al. Frequency, of malignancy in dermatomyositis: analysis of 239 mutual exclusivity and clinical associations of myositis patients from northern China. Oncol Lett 2018;16: autoantibodies in a combined European cohort of 5960–8. idiopathic inflammatory myopathy patients. J Autoimmun 2019;101:48–55. 60 Mugii N, Hasegawa M, Matsushita T et al. Oropharyngeal dysphagia in dermatomyositis: 72 Rogers A, Chung L, Li S, Casciola-Rosen L, Fiorentino associations with clinical and laboratory features DF. Cutaneous and systemic findings associated with including autoantibodies. PLoS One 2016;11: nuclear matrix protein 2 antibodies in adult e0154746. dermatomyositis patients. Arthritis Care Res 2017;69: 1909–14. 61 Fiorentino DF, Chung LS, Christopher-Stine L et al. Most patients with cancer-associated dermatomyositis have 73 Ichimura Y, Matsushita T, Hamaguchi Y et al. Anti-NXP2 antibodies to nuclear matrix protein NXP-2 or autoantibodies in adult patients with idiopathic transcription intermediary factor 1c. Arthritis Rheum inflammatory myopathies: possible association with 2013;65:2954–62. Nov malignancy. Ann Rheum Dis 2012;71:710–3. 62 Li S, Ge Y, Yang H et al. The spectrum and clinical 74 Ishikawa A, Muro Y, Sugiura K, Akiyama M. significance of myositis-specific autoantibodies in Development of an ELISA for detection of autoantibodies Chinese patients with idiopathic inflammatory to nuclear matrix protein 2. Rheumatology (United myopathies. Clin Rheumatol 2019;38:2171–9. Kingdom)2012;51:1181–7. https://academic.oup.com/rheumatology 2627 Alexander G. S. Oldroyd et al. 75 Albayda J, Pinal-Fernandez I, Huang W et al. Antinuclear 85 Leatham H, Schadt C, Chisolm S et al. Evidence matrix protein 2 autoantibodies and edema, muscle supports blind screening for internal malignancy in disease, and malignancy risk in dermatomyositis dermatomyositis: data from 2 large US dermatology patients. Arthritis Care Res 2017;69:1771–6. cohorts. Medicine (Baltimore) 2018;97:e9639. 76 Wang L, Huang L, Yang Y et al. Calcinosis and 86 Huang Z, Gao B, Chen H et al. An efficacy analysis of malignancy are rare in Chinese adult patients with whole-body magnetic resonance imaging in the myositis and nuclear matrix protein 2 antibodies diagnosis and follow-up of polymyositis and identified by an unlabeled immunoprecipitation assay. dermatomyositis. PLoS One 2017;12:e0181069. Clin Rheumatol 2018;37:2731–9. 87 Whitmore SE, Anhalt GJ, Provost TT et al. Serum CA- 77 Marie I, Josse S, Decaux O et al. Comparison of long- 125 screening for ovarian cancer in patients with term outcome between anti-Jo1- and anti-PL7/PL12 dermatomyositis. Gynecol Oncol 1997;65:241–4. positive patients with antisynthetase syndrome. 88 Selva-O’Callaghan A, Grau JM, Gámez-Cenzano C et al. Autoimmun Rev 2012;11:739–45. Conventional cancer screening versus PET/CT in 78 Pinal-Fernandez I, Casal-Dominguez M, Huapaya JA et dermatomyositis/polymyositis. Am J Med 2010;123:558–62. al. A longitudinal cohort study of the anti-synthetase 89 Amoura Z, Duhaut P, Huong DLT et al. Tumor antigen syndrome: increased severity of interstitial lung disease markers for the detection of solid cancers in in black patients and patients with anti-PL7 and anti- inflammatory myopathies. Cancer Epidemiol Biomarkers PL12 autoantibodies. Rheumatology (Oxford) 2017;56: Prev 2005;14:1279–82. 999–1007. 90 Tiniakou E, Pinal-Fernandez I, Lloyd TE et al. More 79 Hamaguchi Y, Fujimoto M, Matsushita T et al. severe disease and slower recovery in younger patients Common and distinct clinical features in adult patients with anti-3-hydroxy-3-methylglutarylcoenzyme A with anti-aminoacyl-tRNA synthetase antibodies: reductase-associated autoimmune myopathy. heterogeneity within the syndrome. PLoS One 2013;8: Rheumatology (Oxford) 2017;56:787–94. e60442. 91 Motegi S, Fujiwara C, Sekiguchi A et al. Clinical value of 80 Allenbach Y, Keraen J, Bouvier AM et al. High risk of 18 F-fluorodeoxyglucose positron emission tomography/ cancer in autoimmune necrotizing myopathies: usefulness computed tomography for interstitial lung disease and of myositis specific antibody. Brain 2016;139:2131–5. myositis in patients with dermatomyositis. J Dermatol 2019; 81 Aggarwal R, Moghadam-Kia S, Lacomis D et al. Anti- 46:213–8. hydroxy-3-methylglutaryl-coenzyme A reductase (anti- 92 Kundrick A, Kirby J, Ba D et al. Positron emission HMGCR) antibody in necrotizing myopathy: treatment tomography costs less to patients than conventional outcomes, cancer risk, and role of autoantibody level. screening for malignancy in dermatomyositis. Semin Scand J Rheumatol 2020;49:405–11. Arthritis Rheum 2019;49:140–4. 82 Hoesly PM, Sluzevich JC, Jambusaria-Pahlajani A et al. 93 Brix G, Lechel U, Glatting G et al. Radiation exposure of Association of antinuclear antibody status with clinical patients undergoing whole-body dual-modality 18F-FDG features and malignancy risk in adult-onset PET/CT examinations. J Nucl Med 2005;46:608–13. dermatomyositis. J Am Acad Dermatol 2019;80:1364–70. 94 Loarce-Martos J, Lilleker JB, Parker M, McHugh N, 83 Kidambi TD, Schmajuk G, Gross AJ et al. Endoscopy is Chinoy H. Polymyositis: is there anything left? A of low yield in the identification of gastrointestinal retrospective diagnostic review from a tertiary myositis neoplasia in patients with dermatomyositis: a cross- centre. Rheumatology 2020; Advance Access published sectional study. World J Gastroenterol 2017;23: 24 December 2020, doi: 10.1093/rheumatology/keaa801. 4788–95. 95 van der Meulen MF, Bronner IM, Hoogendijk JE et al. 84 Maliha PG, Hudson M, Abikhzer G, Singerman J, Probst Polymyositis: an overdiagnosed entity. Neurology 2003; S. 18F-FDG PET/CT versus conventional investigations 61:316–21. for cancer screening in autoimmune inflammatory myopathy in the era of novel myopathy classifications. 96 Amato AA, Griggs RC. Unicorns, dragons, polymyositis, Nucl Med Commun 2019;40:377–82. and other mythological beasts. Neurology 2003;61:288–9. 2628 https://academic.oup.com/rheumatology