www.nature.com/bjc ARTICLE Molecular Diagnostics Poor outcome in hypoxic endometrial carcinoma is related to vascular density Casper Reijnen1,2, Willem Jan van Weelden1, Martijn S. J. P. Arts3, Johan P. Peters3, Paul F. Rijken3, Koen van de Vijver4, Maria Santacana5, Peter Bronsert6, Johan Bulten7, Marc Hirschfeld8,9, Eva Colas10, Antonio Gil-Moreno10,11, Armando Reques12, Gemma Mancebo13, Camilla Krakstad14,15, Jone Trovik14, Ingfrid S. Haldorsen15,16, Jutta Huvila17, Martin Koskas18, Vit Weinberger19, Lubos Minar19, Eva Jandakova20, Marc P. L. M. Snijders2, Saskia van den Berg-van Erp21, Heidi V. N. Küsters-Vandevelde21, Xavier Matias-Guiu5, Frederic Amant22,23, ENITEC-consortium, Leon F. A. G. Massuger1, Johan Bussink3 and Johanna M. A. Pijnenborg1 BACKGROUND: Identification of endometrial carcinoma (EC) patients at high risk of recurrence is lacking. In this study, the prognostic role of hypoxia and angiogenesis was investigated in EC patients. METHODS: Tumour slides from EC patients were stained by immunofluorescence for carbonic anhydrase IX (CAIX) as hypoxic marker and CD34 for assessment of microvessel density (MVD). CAIX expression was determined in epithelial tumour cells, with a cut-off of 1%. MVD was assessed according to the Weidner method. Correlations with disease-specific survival (DSS), disease-free survival (DFS) and distant disease-free survival (DDFS) were calculated using Kaplan–Meier curves and Cox regression analysis. RESULTS: Sixty-three (16.4%) of 385 ECs showed positive CAIX expression with high vascular density. These ECs had a reduced DSS compared to tumours with either hypoxia or high vascular density (log-rank p = 0.002). Multivariable analysis showed that hypoxic tumours with high vascular density had a reduced DSS (hazard ratio [HR] 3.71, p = 0.002), DDFS (HR 2.68, p = 0.009) and a trend for reduced DFS (HR 1.87, p = 0.054). CONCLUSIONS: This study has shown that adverse outcome in hypoxic ECs is seen in the presence of high vascular density, suggesting an important role of angiogenesis in the metastatic process of hypoxic EC. Differential adjuvant treatment might be indicated for these patients. British Journal of Cancer (2019) 120:1037–1044; https://doi.org/10.1038/s41416-019-0461-2 BACKGROUND their vasculature beyond the size of 0.1 mm3, hypoxia may occur.9 Most endometrial carcinoma (EC) patients present with early-stage As a response to chronic hypoxia, tumour cells will activate genes disease and have a favourable outcome. Nevertheless, 15% of all associated with more aggressive phenotype and resistance to patients suffer from recurrent disease and subsequently have a chemotherapy and radiotherapy.10 Hypoxia-inducible factor 1 poor outcome.1–3 Approximately half of these recurrences occur in (HIF-1), formed after heterodimerisation of its subunits HIF-1α and patients primarily diagnosed with low-risk EC.1,4 Improved HIF-1β, plays a key role in this process.11,12 HIF-1 activates identification of patients at high risk for recurrence is crucial to downstream genes that enhance cell survival by maintaining prevent both over- and undertreatment. intracellular pH, stimulating angiogenesis to increase oxygen Hypoxia is known to be an important feature of aggressive EC delivery and switching to anaerobic glycolysis.12,13 More specifi- and drives metastatic potential.5–8 When solid tumours outgrow cally, an important downstream target is carbonic anhydrase 9 1 Department of Obstetrics and Gynaecology, Radboud Institute for Health Sciences, Radboud University Medical Center, Nijmegen, The Netherlands; 2Department of Obstetrics and Gynaecology, Canisius-Wilhelmina Hospital, Nijmegen, The Netherlands; 3Department of Radiation Oncology, Radboud University Medical Center, Nijmegen, The Netherlands; 4Department of Pathology, Ghent University Hospital, Cancer Research Institute Ghent (CRIG), Ghent, Belgium; 5Department of Pathology and Molecular Genetics and Research Laboratory, Hospital Universitari Arnau de Vilanova, University of Lleida, IRBLleida, CIBERONC, Lleida, Spain; 6Institute of Pathology, University Medical Center, Freiburg, Germany; 7Department of Pathology, Radboud University Medical Center, Nijmegen, The Netherlands; 8Department of Obstetrics and Gynecology, University Medical Center, Freiburg, Germany; 9German Cancer Consortium, German Cancer Research Center, Heidelberg, Germany; 10Biomedical Research Group in Gynecology, Vall Hebron Institute of Research, Universitat Autònoma de Barcelona, CIBERONC, Barcelona, Spain; 11Gynecological Department, Vall Hebron University Hospital, CIBERONC, Barcelona, Spain; 12 Pathology Department, Vall Hebron University Hospital, CIBERONC, Barcelona, Spain; 13Department of Obstetrics and Gynecology, Hospital del Mar, PSMARB, Barcelona, Spain; 14 Department of Obstetrics and Gynecology, Haukeland University Hospital, Bergen, Norway; 15Centre for Cancer Biomarkers, Department of Clinical Science, University of Bergen, Bergen, Norway; 16Mohn Medical Imaging and Visualization Centre, Department of Radiology, Haukeland University Hospital, Bergen, Norway; 17Department of Pathology, University of Turku, Turku, Finland; 18Obstetrics and Gynecology Department, Bichat-Claude Bernard Hospital, Paris, France; 19Department of Gynecology and Obstetrics, Faculty of Medicine, Masaryk University, Brno, Czech Republic; 20Institute of Pathology, Faculty of Medicine, Masaryk University, Brno, Czech Republic; 21Department of Pathology, Canisius-Wilhelmina Hospital, Nijmegen, The Netherlands; 22Department of Oncology, KU Leuven, Leuven, Belgium and 23Department of Gynaecologic Oncology, Center for Gynaecologic Oncology, Amsterdam, The Netherlands Correspondence: Casper Reijnen (Casper.reijnen@radboudumc.nl) These authors contributed equally: Casper Reijnen, Willem Jan van Weelden. Received: 5 December 2018 Revised: 1 April 2019 Accepted: 4 April 2019 Published online: 23 April 2019 © Cancer Research UK 2019 Poor outcome in hypoxic endometrial carcinoma is related to vascular. . . C Reijnen et al. 1038 (CA9), whose encoded protein, carbonic anhydrase IX (CAIX), Image analysis regulates intracellular pH by converting carbon dioxide to Tumour slides were analysed using a digital image analysis system carbonic acid.14 By adaptation of tumour cells to a hostile after scanning of the whole slides with the Axio Imager D2 microenvironment, tumour proliferation can commence even in microscope (Carl Zeiss, GmbH, Oberkochen, Germany) using a hypoxic areas.15 Also in normoxic conditions, HIF-1 can be Prior lumen 200 metal halide lamp (Prior Scientific, Rockland, USA), activated; however, downstream activation is present in lesser Axiocam 503 mono 16-bit camera (1936 × 1460 pixel, Carl Zeiss, extent.16,17 In this perspective CAIX expression, one of the key GmbH) and a computer-controlled motorised stage (Carl Zeiss, effector proteins of HIF-1, has been shown to be more specifically GmbH) directed by Zen Pro software (Carl Zeiss, GmbH).26 Each related to hypoxia and poor outcome.18 slide was scanned for three signals: DAPI (all nuclei), Alexa488 Next to maintenance of intracellular pH, stimulation of (CD34) and Cy3 (CAIX), by means of a 10× objective using angiogenesis is an important response to hypoxia. Vascular standardised shutter times for each signal (1, 25 and 50 ms, endothelial growth factor (VEGF), another downstream target of respectively). After scanning, grey-scale images of all three HIF-1, is also correlated with hypoxia and angiogenesis in several recorded signals were used for analysis. cancer types, including EC.19–22 Angiogenesis can be assessed by For analysis of CAIX staining, only membranous expression on microvessel density (MVD) and is prognostically associated with epithelial tumour cells was analysed. Areas of necrosis, large deep myometrial invasion (MI), lymphovascular space invasion vessels and tumour stroma, determined using H&E-stained (LVSI) and poor overall survival in EC.23 Although earlier studies adjacent tumour slides, were therefore manually excluded from suggest correlation between hypoxia, angiogenesis and poor the analysis (i-Vision for Mac; BioVision Technologies, Exton, PA, outcome, the prognostic value has not yet been studied before.5,6 USA). Next, thresholds for segmentation of the fluorescent Therefore, we have investigated the prognostic value of hypoxia signals were interactively set above the background staining for and angiogenesis in EC, assessed with CAIX expression and MVD. each individual marker and adjusted for each sample in order to optimise the signal to background ratio using ImageJ software (Wayne Rasband, National Institute of Mental Health, National METHODS Institutes of Health). An interactively set threshold limits inter- Patients sample variability by correction for differences in immunofluor- 1234567890();,: Data and tumour slides were collected previously for a study escence staining intensity.26,27 The resulting binary images were analysing the value of L1CAM expression in ECs, which included used to calculate the fraction of CAIX (fCAIX) relative to the total ECs from 11 collaborating European Network for Individualised tumour area. To minimise bias of non-specific staining, only Treatment of Endometrial Cancer (ENITEC) centres.24,25 Only cases positive signals exceeding 5 pixels were included. diagnosed by an expert gynaecological pathologist, with com- The MVD was measured according to the Weidner method.28 plete data on treatment and pathological examination and at least In short, surrounding epithelial tumour cells three areas with 36 months of follow-up, were included. Cases with a non- the highest density of vessels were selected by the assessor endometrioid component were categorised as non-endometrioid. (M.A.) using a ×200 magnification. To correct for objects that The 1199 cases included in the original study were randomly exceed the image borders, only objects exceeding the left selected using SPSS version 22 (SPSS IBM, New York, NY, USA), and upper border were included. To minimise bias of non- resulting in a database of 403 patients for the present study. These specific staining, only positive signals exceeding 2 pixels were cases were not statistically different from the original cases for the included. most important baseline characteristics. CAIX expression was considered positive when the fCAIX was above 1%.29,30 The MVD was dichotomised over the median. A Tissue and staining representative example of CAIX and MVD staining is shown in Four micrometre sections, derived from formalin-fixed, paraffin- Fig. 1. embedded ECs were used to visualise CAIX and blood vessels. Sections were mounted on Superfrost slides (Menzel–Gläser). Statistical analyses Slides were deparaffinated in Histochoise (VWR H103-4L) and Clinicopathological differences between subgroups were com- rehydrated (graded ethanol: 100–96–70% and de-ionised water). pared with the χ2 and Fisher’s exact tests for categorical data and Next, citrate buffer antigen retrieval was performed for 30 min the Mann–Whitney U test for continuous variables. (Target retrieval solution 10×, pH 6 citrate, Dako Cytomation, Kaplan–Meier curves were constructed for disease-specific 96 °C). Prior to incubation with the primary antibodies, sections survival (DSS), disease-free survival (DFS) and distant-DFS (DDFS). were blocked with 5% normal goat serum (Jackson ImmunoR- The association between CAIX and MVD and DSS, DFS and DDFS esearch) in primary antibody diluent (PAD, BIORAD BUF014), 30 was determined using Cox regression analysis. DSS was calculated min at room temperature. Thereafter, sections were co-stained for from the date of primary treatment to the date of death caused by CAIX (Novus Biologicals NB100-417, 1:100) and vessels (CD34, the disease or, for surviving patients, to the date of the last follow- ABCAM ab8536, 1:300), 60 min at 37 °C. Secondary incubation was up. DFS and DDFS were defined as the length of follow-up, after performed using CyTM3 Fab Fragment Goat Anti-Rabbit immu- completion of the primary treatment, during which women noglobulin G (IgG) (H + L) polyclonal IgG (Jackson ImmunoRe- survived without any clinical sign of (distant) disease recurrence. search 111-167-003) for CAIX and CF®488a Goat anti-Mouse IgG (H Distant recurrence included metastases in distant organs and + L), F(ab′)2 fragment polyclonal IgG (Biotium CF488A) for the para-aortic lymph nodes. Features identified by univariable vessels, 60 min at 37 °C. All antibodies were diluted in PAD. In regression analysis with p < 0.20 were used for multivariable between stainings, sections were rinsed with phosphate-buffered regression analysis. LVSI was coded as negative in case of missing saline (JT Baker 4391.9010). 4′,6-Diamidino-2-phenylindole (DAPI) data (n = 108) since only substantial LVSI was recently reported as (Santa Cruz Biotechnology AB-17.0097) was used as a counterstain relevant for prognosis of EC. If LVSI was not reported in the to stain all nuclei, and finally the sections were mounted with pathological report, it was therefore assumed that LVSI was Fluoromount W (Serva 21634.01). Haematoxylin and eosin staining absent.31,32 P values <0.05 were considered to indicate a was used for morphological evaluation. CAIX expression was significant difference. SPSS version 25 (SPSS IBM, New York, NY, scored as the fraction of epithelial tumour cells with positive USA) statistical software was used to perform the statistical membranous staining. analyses. Poor outcome in hypoxic endometrial carcinoma is related to vascular. . . C Reijnen et al. 1039 a e b f c g d h Fig. 1 Example of carbonic anhydrase IX (CAIX) and CD34 staining in endometrial cancer. a Nuclear 4′,6-diamidino-2-phenylindole (DAPI) staining (blue) for visualisation of tumour nuclei. b CAIX staining (green) adjusted for total tumour area, meaning that only epithelial tumour cells were included in the analysis: other tissue, including stroma, necrosis and vasculature has manually been removed and is coloured black by the analysis software (see “Methods” section). c CD34 staining of vasculature (red) with three hotspots according to the Weidner method, marked with the interrupted lines. d Combined CAIX staining and CD34 staining. e–h represent representative high-magnification images of the boxed areas in a, b, c and d, respectively. Scale bar = 0.5 mm Poor outcome in hypoxic endometrial carcinoma is related to vascular. . . C Reijnen et al. 1040 RESULTS CAIX expression and MVD Patients A total of 63 carcinomas (16.4%) showed a positive membranous After staining for CAIX and CD34, 18 of the 403 patients were epithelial CAIX expression and high degrees of vascular density, excluded due to insufficient tumour tissue (n = 9) and excess of non- defined as a MVD above the median (Table 1). CAIX expression specific background staining (n = 9). Clinicopathological character- with high vascular density was correlated with non-endometrioid istics of the 385 patients included for analysis are shown in Table 1. histology (7.9% vs. 2.5%, p = 0.028), but not with other Overall, the median age was 64 years and the median follow-up time clinicopathological features. Patients with CAIX-positive ECs and was 58 months. Of all patients alive at the end of follow-up, 99% had high vascular density experienced more recurrences (22.2% vs. a follow-up of at least 36 months. A total of 67 patients (17.4%) were 10.2%, p = 0.008) and specifically more distant recurrences (19.0% diagnosed with high-grade EC, including 13 non-endometrioid vs. 5.9%, p < 0.001), as well as higher overall mortality (22.2% vs. endometrial carcinomas (NEECs) (3.4%). In total, 106 (27%) EC 11.8%, p = 0.027) and EC-related mortality (17.5% vs. 3.7%, p < patients had positive CAIX expression. Forty-seven patients (12.2%) 0.001) (Table 1). recurred and 21 patients (5.5%) died due to the disease. Of all the Figure 2 shows that CAIX expression with high vascular density patients with recurrence, 14 (3.6%) had a local recurrence, 16 (4.2%) was associated with a worse DSS compared to CAIX expression a regional recurrence and 31 (8.1%) a distant recurrence. with low vascular density and negative CAIX expression (p = Table 1. Baseline characteristics of all included patients, associated with CAIX combined with vascular density Variable All (n = 385) Not CAIX >1% and high vascular density CAIX >1% and high vascular density P* (n = 322) (n = 63) Age (years)a 64.0 (34.0–89.0) 64.3 (34.0–89.0) 63.5 (43.0–88.0) 0.634 Follow-up (months)a 65.1 (0.0–156.0) 63.8 (0.0–156.0) 71.6 (0–148.0) 0.125 Grade Low 318 (82.6) 270 (83.9) 48 (76.2) 0.142 High 67 (17.4) 52 (16.1) 15 (23.8) Histology EEC 372 (96.6) 314 (97.5) 58 (92.1) 0.028 NEEC 13 (3.4) 8 (2.5) 5 (7.9) FIGO stage I–II 363 (93.8) 305 (94.7) 58 (92.1) 0.406 III–IV 22 (5.7) 17 (5.3) 5 (7.9) Myometrial invasion <50% 258 (67.2) 218 (67.9) 40 (63.5) 0.494 ≥50% 126 (32.8) 103 (32.1) 23 (36.5) LVSIb No 243 (62.8) 200 (90.1) 43 (84.3) 0.234 Yes 30 (7.8) 22 (9.9) 8 (15.7) Lymph nodesc No metastasis 263 (68.0) 223 (95.7) 40 (95.2) 0.891 Metastasis 12 (3.1) 10 (4.3) 2 (4.8) Adjuvant treatment No 154 (40.0) 143 (44.8) 22 (34.9) Radiotherapy 200 (51.9) 161 (50.0) 39 (61.9) 0.084 Chemotherapy 31 (8.1) 18 (5.6) 2 (3.2) 0.429 Recurrence No 338 (87.8) 289 (89.8) 49 (77.8) Yes 47 (12.2) 33 (10.2) 14 (22.2) 0.008 Local 14 (3.6) 12 (3.7) 2 (3.2) 0.830 Regional 16 (4.2) 13 (4.0) 3 (4.8) 0.792 Distant 31 (8.1) 19 (5.9) 12 (19.0) <0.001 Death No 335 (87.0) 284 (88.2) 49 (77.8) Yes 50 (13.0) 38 (11.8) 14 (22.2) 0.027 EC-related 21 (5.5) 12 (3.7) 11 (17.5) <0.001 CAIX carbonic anhydrase IX, EEC endometrioid endometrial carcinoma, NEEC non-endometrioid endometrial carcinoma, FIGO International Federation of Gynaecology and Obstetrics, LVSI lymphovascular space invasion, EC endometrial carcinoma *P value of the Mann–Whitney U test for continuous, and χ2 test and Fisher’s exact for categorical variables a Median values (range) b Based on 273 patients c Based on 275 patients Poor outcome in hypoxic endometrial carcinoma is related to vascular. . . C Reijnen et al. 1041 0.002). Interestingly, CAIX-positive ECs with low vascular density with deep MI, but not with other clinicopathological factors. In had a similar outcome as CAIX-negative ECs. Univariable Cox multivariable analysis, high MVD remained an independent regression analysis revealed that age, CAIX expression with high predictor of reduced DSS (HR 2.92, 95% CI 1.13–7.54, p = 0.027) vascular density, myocardial infarction (MI), International Federa- (Supplementary Tables 3 and 4). Continuous scoring of CAIX tion of Gynaecology and Obstetrics (FIGO) stage, grade and expression showed a significant correlation with DSS as well (data lymphovascular space invasion (LVSI) were significantly associated not shown). with DSS (Fig. 3). In multivariable analysis, high age, CAIX expression with high vascular density and tumour grade 3 remained significantly associated with reduced DSS, with CAIX DISCUSSION and MVD as the most significant parameter (hazard ratio [HR] 3.71, In the present study, we have investigated the prognostic value of 95% confidence interval (CI) 1.59–8.63, p = 0.002). angiogenesis and hypoxia, assessed with MVD and CAIX expres- Multivariable analysis showed that age, FIGO stage and LVSI sion. We hypothesised that angiogenesis would facilitate haema- were significantly associated with DFS. CAIX expression with high togenous spread of hypoxic tumour cells with subsequent poor vascular density was nearly significant (HR 1.87, 95% CI 0.99–3.55, clinical outcome. Additionally, we assumed that this would p = 0.054, Fig. 4). Multivariable analysis for DDFS showed that LVSI specifically be facilitated in hypoxic ECs, because of activation of and CAIX expression with high vascular density was significantly intracellular pathways that induce an aggressive and metastatic associated with a reduced DDFS (CAIX and MVD: HR 2.68, 95% CI phenotype. We have shown that CAIX expression with high 1.27–5.65, p = 0.009, Fig. 5). vascular density is associated with reduced DSS and DDFS. Interestingly, CAIX-positive ECs with low vascular density had a Individual contribution of CAIX and MVD similar outcome as CAIX-negative ECs. Finally, multivariable Positive CAIX expression was associated with high tumour grade, analyses for CAIX expression and vascular density showed that non-endometrioid histology, higher median MVD and treatment both were independent prognostic markers as well. with radiotherapy. In multivariable analysis, CAIX and grade were This is the largest study to date studying CAIX in EC. In contrast significantly associated with DSS (HR 2.45, 95% CI 1.05–5.73, p = to previous studies in EC, we did find significant correlations 0.039) (Supplementary Tables 1 and 2). High MVD was correlated between CAIX expression and poor outcome, especially in case of high vascular density. Seeber et al.29 included 93 patients and Disease-specific survival found CAIX expression in 76% of ECs.29 In this study, no 1.0 correlation between CAIX expression and outcome was found; however, small sample size and different cut-off value (all degrees of positive staining were regarded as positive) could explain 0.8 why no correlation was found. Similarly, Pijnenborg et al.22 Cumulative survival investigated CAIX expression in 59 ECs and did not find a 0.6 correlation. Again, possibly this study was underpowered due to a limited sample size and low number of distant recurrences. Also, 0.4 differences in study design (case–control study) hamper valid comparison with our results. In other cancer types, including 0.2 CAIX-negative, low angiogenesis breast carcinoma, hepatocellular carcinoma, cervical carcinoma CAIX-negative, high angiogenesis CAIX-positive, low angiogenesis and renal cell carcinoma, CAIX expression is associated with poor CAIX-positive, high angiogenesis prognosis.33–37 More specifically, increased distant failure was seen log-rank: p = 0.002 0.0 in several solid tumour types with positive CAIX expression.37,38 0 20 40 60 80 100 120 140 The metastatic process is a complex step-wise process, Follow-up (months) including acquisition of an aggressive phenotype, invasion in Fig. 2 Disease-specific survival (DSS) by carbonic anhydrase IX surrounding tissues and blood vessels, survival in the circulation (CAIX) expression combined with degree of angiogenesis. Log-rank with subsequent extravasation and colonisation in new organs.39 test was used to compare groups Hypoxia and subsequent neoangiogenesis will intervene with Disease-specific survival Age (years) <65 – >65 3.74 (1.48–9.49) p = 0.005 2.78 (1.08–7.17) p = 0.034 CAIX&MVD Negative-Positive 4.28 (1.88–9.76) p < 0.001 3.71 (1.59–8.63) p = 0.002 Myometrial invasion <50% – >50% 3.22 (1.39–7.43) p = 0.006 1.41 (0.53–3.75) p = 0.492 FIGO-stage I/II-III/IV 7.35 (2.86–18.88) p < 0.001 3.15 (0.90–10.96) p = 0.072 Grade 1/2–3 6.39 (2.81–14.53) p < 0.001 2.75 (1.03–7.31) p = 0.043 LVSI No-Yes 5.58 (2.30–13.58) p < 0.001 1.53 (0.47–5.00) p = 0.485 1 6 11 16 Hazard ratio (95% CI) Fig. 3 Univariable and multivariable Cox regression analysis of clinicopathological parameters including carbonic anhydrase IX (CAIX) combined with vascular density for disease-specific survival (DSS). The hazard ratios with 95% confidence intervals are depicted by the black line. All risk factors significantly associated with DSS in univariable analysis were included in the multivariable Cox regression analysis, depicted by the grey lines Poor outcome in hypoxic endometrial carcinoma is related to vascular. . . C Reijnen et al. 1042 Disease-free survival Age (years) <65 – >65 3.21 (1.72–6.00) p < 0.001 2.76 (1.46–5.23) p = 0.002 CAIX&MVD Negative-Positive 2.05 (1.09–3.86) p = 0.025 1.87 (0.99–3.55) p = 0.054 Myometrial invasion <50% – >50% 2.43 (1.37–4.31) p = 0.002 1.11 (0.56–2.20) p = 0.763 FIGO-stage I/II-III/IV 6.44 (3.26–12.71) p < 0.001 2.53 (1.03–6.25) p = 0.044 Grade 1/2–3 3.99 (2.22–7.16) p < 0.001 1.69 (0.81–3.52) p = 0.160 LVSI No-Yes 6.10 (3.26–11.4) p < 0.001 2.85 (1.24–6.58) p = 0.014 0 5 10 15 20 Hazard ratio (95% CI) Fig. 4 Univariable and multivariable Cox regression analysis of clinicopathological parameters including carbonic anhydrase IX (CAIX) combined with vascular density for disease-free survival (DFS). All risk factors significantly associated with DFS in univariable analysis were included in the multivariable Cox regression analysis, depicted by the grey lines Distant disease-free survival Age (years) <65 – >65 2.87 (1.35–6.10) p = 0.006 2.28 (1.06–4.92) p = 0.036 CAIX&MVD Negative-Positive 3.06 (1.48–6.33) p = 0.003 2.68 (1.27–5.65) p = 0.009 Myometrial invasion <50% – >50% 3.87 (1.85–8.07) p < 0.001 1.66 (0.70–3.91) p = 0.250 FIGO-stage I/II-III/IV 7.32 (3.26–16.45) p < 0.001 1.84 (0.66–5.15) p = 0.245 Grade 1/2–3 6.31 (3.11–12.80) p < 0.001 2.25 (0.93–5.48) p = 0.074 LVSI No-Yes 9.09 (4.40–18.74) p < 0.001 3.69 (1.43–9.56) p = 0.007 0 5 10 15 20 Hazard ratio (95% CI) Fig. 5 Univariable and multivariable Cox regression analysis of clinicopathological parameters including carbonic anhydrase IX (CAIX) combined with vascular density for distant disease-free survival (DDFS). All risk factors significantly associated with DDFS in univariable analysis were included in the multivariable Cox regression analysis, depicted by the grey lines several steps of this process, including promoting tumour cell observation that CAIX-positive ECs with high vascular density survival by acquisition of a malignant phenotype and increased did not have more lymph node metastasis or local recurrences, invasion in blood vessels.9 but instead have more distant recurrences, could support the role A recent meta-analysis has shown that high MVD was of angiogenesis in the haematogenous rather than the lympho- associated with several poor prognostic variables, including deep genic metastatic process. MI, positive LVSI and poor outcome in EC, although heterogeneity The obvious strengths of this study are the inclusion of a large due to differences in used antibodies and cut-off values hampers and representative cohort of EC patients within the ENITEC interpretation of these results.23 Biologically, intratumoural network and the objective and reproducible measurement of neoangiogenesis in response to hypoxia will promote the CAIX and MVD using digital imaging analyses. However, there formation of vasculature with high degrees of permeability and are some limitations that need to be addressed. Due to the potential for rapid growth.40 Our hypothesis that CAIX expression retrospective nature of the study, there were missing values, with high degrees of vascular density would be associated with specifically for LVSI and lymph node metastasis. Substantial LVSI unfavourable prognostic features and poor outcome was based is a stronger predictor for prognosis of EC compared to both on the facilitation of haematogenous spread in areas with moderate LVSI. Also, LVSI is not routinely reported in the high angiogenesis and on the aggressive biological behaviour of pathologic report at all centres. Therefore, we assumed that if tumour cells after hypoxia.41,42 HIF-1α is stabilised and accumu- substantial LVSI was present, it was reported, and if LVSI was not lates under hypoxia, and activates transcription of numerous reported, no substantial LVSI was present.31 Missing cases were genes involved in angiogenesis, proliferation and pH regulation therefore coded as negative for LVSI. Separate analyses of (VEGF, CAIX, glucose transporter-1).9 Our hypothesis was sup- patients with available LVSI status did not alter the results of the ported by the fact that ECs with positive CAIX expression and high primary outcome (data not shown). Another general limitation vascular density had a decreased DSS compared to ECs with only in interpretation of CAIX and MVD is the lack of standardised one or none of both features. This observation supports the criteria in the current literature, which hampers comparison of complex interplay underlying the metastatic processes. The previous studies and this study.29,30 However, the applied digital Poor outcome in hypoxic endometrial carcinoma is related to vascular. . . C Reijnen et al. 1043 techniques in this study enable objective and reproducible Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims analyses without the need for extensive pathological expertise. in published maps and institutional affiliations. With the integration of digital pathology into clinical practice, comparison of future studies with our results might be easier.43,44 Although widely used to quantify MVD, CD34 is REFERENCES known to also identify lymph vessels and stem cell populations, 1. Creasman, W. T., Morrow, C. P., Bundy, B. N., Homesley, H. D., Graham, J. E., Heller, which theoretically could have led to an overestimation of our P. B. Surgical pathologic spread patterns of endometrial cancer. A Gynecologic Oncology Group Study. Cancer 60(Suppl.), 2035–2041 (1987). results. On the other hand, other antibodies, for example, CD31, 2. Bokhman, J. V. Two pathogenetic types of endometrial carcinoma. Gynecol. Oncol. also carry the risk of aspecific staining. Compared to CD31, 15, 10–17 (1983). CD34 staining is known to have stronger reactivity with 3. Nout, R. A., Smit, V. T., Putter, H., Jurgenliemk-Schulz, I. M., Jobsen, J. J., Lutgens, L. endothelial cells, resulting in a lower risk of staining failure.45 C. et al. Vaginal brachytherapy versus pelvic external beam radiotherapy for Finally, generalisability to non-endometrioid subtypes can be patients with endometrial cancer of high-intermediate risk (PORTEC-2): an open- questioned, as they comprise only 3.4% of the entire cohort. label, non-inferiority, randomised trial. Lancet 375, 816–823 (2010). More research focused on this specific subgroup could help to 4. Trovik, J., Wik, E., Werner, H. M., Krakstad, C., Helland, H., Vandenput, I. et al. strengthen these results. Hormone receptor loss in endometrial carcinoma curettage predicts lymph node This study identifies a group of patients with a poor DSS and metastasis and poor outcome in prospective multicentre trial. Eur. J. Cancer 49, 3431–3441 (2013). DDFS based on CAIX and MVD. Given the increased risk 5. Berg, A., Fasmer, K. E., Mauland, K. K., Ytre-Hauge, S., Hoivik, E. A., Husby, J. A. et al. of distant metastases, differential adjuvant treatment for these Tissue and imaging biomarkers for hypoxia predict poor outcome in endometrial ECs could be explored either in the form of chemotherapy or, in cancer. Oncotarget 7, 69844–69856 (2016). the future, targeted therapies directed against angiogenesis. 6. Haldorsen, I. S., Stefansson, I., Gruner, R., Husby, J. A., Magnussen, I. J., Werner, H. Because of the focal character of CAIX expression in the tumour M. et al. Increased microvascular proliferation is negatively correlated to tumour tissue, performing the analysis on preoperative biopsies blood flow and is associated with unfavourable outcome in endometrial carci- might be challenging, but visualisation of hypoxia and nomas. Br. J. Cancer 110, 107–114 (2014). angiogenesis on FDG-PET/CT (fluorodeoxyglucose-positron 7. Gilkes, D. M., Semenza, G. L. & Wirtz, D. Hypoxia and the extracellular matrix: emission tomography/computed tomography) scan and mag- drivers of tumour metastasis. Nat. Rev. Cancer 14, 430–439 (2014). 8. Bredholt, G., Mannelqvist, M., Stefansson, I. M., Birkeland, E., Bo, T. H., Oyan, A. M. netic resonance imaging could be an alternative, as Berg et al.5 et al. Tumor necrosis is an important hallmark of aggressive endometrial cancer showed recently. and associates with hypoxia, angiogenesis and inflammation responses. Onco- In summary, we have found that CAIX expression and high target 6, 39676–39691 (2015). vascular density are prognostic markers for decreased survival in 9. Span, P. N. & Bussink, J. Biology of hypoxia. Semin. Nucl. Med 45, 101–109 (2015). EC. Combining these two markers revealed that ECs with positive 10. Rademakers, S. E., Span, P. N., Kaanders, J. H., Sweep, F. C., van der Kogel, A. J. & CAIX expression and high vascular density have an impaired Bussink, J. Molecular aspects of tumour hypoxia. Mol. Oncol. 2, 41–53 (2008). outcome compared to ECs that have only one or none of both 11. Jiang, B. H., Semenza, G. L., Bauer, C. & Marti, H. H. Hypoxia-inducible factor 1 features. These patients experienced more distant recurrences, levels vary exponentially over a physiologically relevant range of O2 tension. Am. and therefore differential adjuvant treatment for these tumours J. Physiol. 271(Part 1), C1172–C1180 (1996). 12. Semenza, G. L. HIF-1: mediator of physiological and pathophysiological responses should be explored. to hypoxia. J. Appl. Physiol. (1985) 88, 1474–1480 (2000). 13. Ratcliffe, P. J., O’Rourke, J. F., Maxwell, P. H. & Pugh, C. W. Oxygen sensing, hypoxia-inducible factor-1 and the regulation of mammalian gene expression. J. AUTHOR CONTRIBUTIONS Exp. Biol. 201(Part 8), 1153–1162 (1998). All authors contributed to the manuscript. C.R., W.J.v.W., M.S.J.P.A., M.P.L.M.S., H.V.N. 14. Ivanov, S. V., Kuzmin, I., Wei, M. H., Pack, S., Geil, L., Johnson, B. E. et al. Down- K.-V., L.F.A.G.M., J.B. and J.M.A.P. contributed to the study design. C.R., W.J.v.W., M.S.J. regulation of transmembrane carbonic anhydrases in renal cell carcinoma cell P.A., J.P.P., P.F.R., K.v.d.V., M.S., P.B., J. B., M.H., E.C., A.G.-M., A.R., G.M., C.K., J.T., I.S.H., J. lines by wild-type von Hippel-Lindau transgenes. Proc. Natl Acad. Sci. USA 95, H., M.K., V.W., L.M., E.J., M.P.L.M.S., S.v.d.B.-v.E., H.V.N.K.-V., X.M.-G., F.A., L.F.A.G.M., J.B., 12596–12601 (1998). J.M.A.P. contributed to data collection. C.R., W.J.v.W., M.S.J.P.A., J.P.P., P.F.R, M.P.L.M.S., 15. Blancher, C. & Harris, A. L. The molecular basis of the hypoxia response pathway: H.V.N.K.-V., L.F.A.G.M., J.B. and J.M.A.P. contributed to the data analysis. C.R., W.J.v.W., tumour hypoxia as a therapy target. Cancer Metastas. Rev. 17, 187–194 (1998). M.S.J.P.A., J.P.P., P.F.R, M.P.L.M.S., H.V.N.K.-V., L.F.A.G.M., J.B and J.M.A.P contributed to 16. Semenza, G. L. Defining the role of hypoxia-inducible factor 1 in cancer biology writing of the draft version of the manuscript. C.R., W.J.v.W., M.S.J.P.A., J.P.P., P.F.R., K.v. and therapeutics. Oncogene 29, 625–634 (2010). d.V., M.S., P.B., J.B., M.H., E.C., A.G.-M., A.R., G.M., C.K., J.T., I.S.H., J.H., M.K., V.W., L.M., E.J., 17. Vleugel, M. M., Greijer, A. E., Shvarts, A., van der Groep, P., van Berkel, M., Aar- M.P.L.M.S., S.v.d.B.-v.E., H.V.N.K.-V., X.M.-G., F.A., L.F.A.G.M., J.B. and J.M.A.P. have read, bodem, Y. et al. Differential prognostic impact of hypoxia induced and diffuse revised and have approved the final version of the manuscript being submitted for HIF-1alpha expression in invasive breast cancer. J. Clin. Pathol. 58, 172–177 peer review. (2005). 18. Bussink, J., Kaanders, J. H. & van der Kogel, A. J. Tumor hypoxia at the micro- regional level: clinical relevance and predictive value of exogenous and endo- ADDITIONAL INFORMATION genous hypoxic cell markers. Radiother. Oncol. 67, 3–15 (2003). Supplementary information is available for this paper at https://doi.org/10.1038/ 19. Leek, R. D., Landers, R. J., Harris, A. L. & Lewis, C. E. Necrosis correlates with high s41416-019-0461-2. vascular density and focal macrophage infiltration in invasive carcinoma of the breast. Br. J. Cancer 79, 991–995 (1999). Competing interests: The authors declare no competing interests. 20. Kato, T., Kimura, T., Miyakawa, R., Tanaka, S., Fujii, A., Yamamoto, K. et al. Clin- icopathologic study of angiogenesis in Japanese patients with breast cancer. World J. Surg. 21, 49–56 (1997). Ethical approval: This study was performed in accordance to the Declaration of 21. Yu, S., Sun, J., Zhang, J., Xu, X., Li, H., Shan, B. et al. Aberrant expression and Helsinki and was approved by the Institutional Review Board at the Radboud association of VEGF and Dll4/Notch pathway molecules under hypoxia in patients University Medical Centre. with lung cancer. Histol. Histopathol. 28, 277–284 (2013). 22. Pijnenborg, J. M., Wijnakker, M., Hagelstein, J., Delvoux, B., Groothuis, P. G. Data availability: The datasets used during the current study can be made available Hypoxia contributes to development of recurrent endometrial carcinoma. Int. J. from the corresponding author on reasonable request. Gynecol. Cancer 17, 897–904 (2007). 23. Wang, J. Z., Xiong, Y. J., Man, G. C. W., Chen, X. Y., Kwong, J., Wang, C. C. Clin- Note: This work is published under the standard license to publish agreement. After icopathological and prognostic significance of blood microvessel density in 12 months the work will become freely available and the license terms will switch to endometrial cancer: a meta-analysis and subgroup analysis. Arch. Gynecol. Obstet. a Creative Commons Attribution 4.0 International (CC BY 4.0). 297, 731–740 (2018). Poor outcome in hypoxic endometrial carcinoma is related to vascular. . . C Reijnen et al. 1044 24. van der Putten, L. J., Visser, N. C., van de Vijver, K., Santacana, M., Bronsert, P., 34. Hussain, S. A., Ganesan, R., Reynolds, G., Gross, L., Stevens, A., Pastorek, J. et al. Bulten, J. et al. L1CAM expression in endometrial carcinomas: an ENITEC colla- Hypoxia-regulated carbonic anhydrase IX expression is associated with poor boration study. Br. J. Cancer 115, 716–724 (2016). survival in patients with invasive breast cancer. Br. J. Cancer 96, 104–109 (2007). 25. van der Putten, L. J. M., Visser, N. C. M., van de Vijver, K., Santacana, M., Bronsert, 35. Cetin, B., Gonul, I. I., Gumusay, O., Bilgetekin, I., Algin, E., Ozet, A. et al. Carbonic P., Bulten, J. et al. Added value of estrogen receptor, progesterone receptor, and anhydrase IX is a prognostic biomarker in glioblastoma multiforme. Neuro- L1 cell adhesion molecule expression to histology-based endometrial carcinoma pathology. https://doi.org/10.1111/neup.12485 (2018). recurrence prediction models: an ENITEC Collaboration Study. Int. J. Gynecol. 36. Finkelmeier, F., Canli, O., Peiffer, K. H., Walter, D., Tal, A., Koch, C. et al. Circulating Cancer. 28, 514–523 (2018). hypoxia marker carbonic anhydrase IX (CA9) in patients with hepatocellular 26. Rijken, P. F., Bernsen, H. J. & van der Kogel, A. J. Application of an image analysis carcinoma and patients with cirrhosis. PLoS ONE 13, e0200855 (2018). system to the quantitation of tumor perfusion and vascularity in human glioma 37. Kirkpatrick, J. P., Rabbani, Z. N., Bentley, R. C., Hardee, M. E., Karol, S., Meyer, J. xenografts. Micro. Res. 50, 141–153 (1995). et al. Elevated CAIX expression is associated with an increased risk of distant 27. Stegeman, H., Rademakers, S. E., Span, P. N., Takes, R. P., van der Kogel, A. J., failure in early-stage cervical cancer. Biomark Insights 3, 45–55 (2008). Kaanders, J. H. et al. Hypoxia, metabolism, and growth factor signaling in head 38. Chia, S. K., Wykoff, C. C., Watson, P. H., Han, C., Leek, R. D., Pastorek, J. et al. and neck squamous cell carcinoma: correlation between primary and xenograft Prognostic significance of a novel hypoxia-regulated marker, carbonic anhydrase tumors. Head. Neck 36, 1288–1295 (2014). IX, in invasive breast carcinoma. J. Clin. Oncol. 19, 3660–3668 (2001). 28. Weidner, N., Semple, J. P., Welch, W. R. & Folkman, J. Tumor angiogenesis and 39. Chang, J. & Erler, J. Hypoxia-mediated metastasis. Adv. Exp. Med. Biol. 772, 55–81 metastasis—correlation in invasive breast carcinoma. N. Engl. J. Med. 324, 1–8 (2014). (1991). 40. Baluk, P., Morikawa, S., Haskell, A., Mancuso, M. & McDonald, D. M. Abnormalities 29. Seeber, L. M., Horree, N., van der Groep, P., van der Wall, E., Verheijen, R. H. & van of basement membrane on blood vessels and endothelial sprouts in tumors. Am. Diest, P. J. Necrosis related HIF-1alpha expression predicts prognosis in patients J. Pathol. 163, 1801–1815 (2003). with endometrioid endometrial carcinoma. BMC Cancer 10, 307 (2010). 41. Yeo, C. D., Kang, N., Choi, S. Y., Kim, B. N., Park, C. K., Kim, J. W. et al. The role of 30. Sadlecki, P., Bodnar, M., Grabiec, M., Marszalek, A., Walentowicz, P., Sokup, A. et al. hypoxia on the acquisition of epithelial-mesenchymal transition and cancer The role of hypoxia-inducible factor-1 alpha, glucose transporter-1, (GLUT-1) and stemness: a possible link to epigenetic regulation. Korean J. Intern. Med. 32, carbon anhydrase IX in endometrial cancer patients. Biomed. Res. Int. 2014, 589–599 (2017). 616850 (2014). 42. Bao, B., Azmi, A. S., Ali, S., Ahmad, A., Li, Y., Banerjee, S. et al. The biological kinship 31. Bosse, T., Peters, E. E., Creutzberg, C. L., Jurgenliemk-Schulz, I. M., Jobsen, J. J., of hypoxia with CSC and EMT and their relationship with deregulated expression Mens, J. W. et al. Substantial lymph-vascular space invasion (LVSI) is a significant of miRNAs and tumor aggressiveness. Biochim. Biophys. Acta 1826, 272–296 risk factor for recurrence in endometrial cancer—a pooled analysis of PORTEC 1 (2012). and 2 trials. Eur. J. Cancer (Oxford, England: 1990) 51, 1742–1750 (2015). 43. Lloyd, M. C., Monaco, J. P. & Bui, M. M. Image analysis in surgical pathology. Surg. 32. Koskas, M., Bassot, K., Graesslin, O., Aristizabal, P., Barranger, E., Clavel-Chapelon, Pathol. Clin. 9, 329–337 (2016). F. et al. Impact of lymphovascular space invasion on a nomogram for predicting 44. Bueno, G., Fernandez-Carrobles, M. M., Deniz, O. & Garcia-Rojo, M. New lymph node metastasis in endometrial cancer. Gynecol. Oncol. 129, 292–297 trends of emerging technologies in digital pathology. Pathobiology 83, 61–69 (2013). (2016). 33. Span, P. N., Bussink, J., Manders, P., Beex, L. V. & Sweep, C. G. Carbonic anhydrase- 45. Uzzan, B., Nicolas, P., Cucherat, M. & Perret, G. Y. Microvessel density as a prog- 9 expression levels and prognosis in human breast cancer: association with nostic factor in women with breast cancer: a systematic review of the literature treatment outcome. Br. J. Cancer 89, 271–276 (2003). and meta-analysis. Cancer Res 64, 2941–2955 (2004)..