f4547c7945f9d69aa248cdb15deb6f50 270279.pdf 38ce9182a69650d7051b571cc3cc906ca80e1696 270279.pdf b360d2d6455b36e93a19bbd328aa588dd545873d517339b96313f78c4942f09e 270279.pdf Title: A Machine Learning Predictive Model for Post-Ureteroscopy Urosepsis Needing Intensive Care Unit Admission: A Case–Control YAU Endourology Study from Nine European Centres Subject: Introduction: With the rise in the use of ureteroscopy and laser stone lithotripsy (URSL), a proportionate increase in the risk of post-procedural urosepsis has also been observed. The aims of our paper were to analyse the predictors for severe urosepsis using a machine learning model (ML) in patients that needed intensive care unit (ICU) admission and to make comparisons with a matched cohort. Methods: A retrospective study was conducted across nine high-volume endourology European centres for all patients who underwent URSL and subsequently needed ICU admission for urosepsis (Group A). This was matched by patients with URSL without urosepsis (Group B). Statistical analysis was performed with ‘R statistical software’ using the ‘randomforests’ package. The data were segregated at random into a 70% training set and a 30% test set using the ‘sample’ command. A random forests ML model was then built with n = 300 trees, with the test set used for internal validation. Diagnostic accuracy statistics were generated using the ‘caret’ package. Results: A total of 114 patients were included (57 in each group) with a mean age of 60 16 years and a male:female ratio of 1:1.19. The ML model correctly predicted risk of sepsis in 14/17 (82%) cases (Group A) and predicted those without urosepsis for 12/15 (80%) controls (Group B), whilst overall it also discriminated between the two groups predicting both those with and without sepsis. Our model accuracy was 81.3% (95%, CI: 63.7–92.8%), sensitivity = 0.80, specificity = 0.82 and area under the curve = 0.89. Predictive values most commonly accounting for nodal points in the trees were a large proximal stone location, long stent time, large stone size and long operative time. Conclusion: Urosepsis after endourological procedures remains one of the main reasons for ICU admission. Risk factors for urosepsis are reasonably accurately predicted by our innovative ML model. Focusing on these risk factors can allow one to create predictive strategies to minimise post-operative morbidity. Keywords: kidney stones; urosepsis; ureteroscopy; laser lithotripsy; urolithiasis; nephrolithiasis; kidney calculi; predictor factors Author: Amelia Pietropaolo, Robert M. Geraghty, Rajan Veeratterapillay, Alistair Rogers, Panagiotis Kallidonis, Luca Villa, Luca Boeri, Emanuele Montanari, Gokhan Atis, Esteban Emiliani, Tarik Emre Sener, Feras Al Jaafari, John Fitzpatrick, Matthew Shaw, Chris Harding and Bhaskar K. Somani Creator: LaTeX with hyperref Producer: pdfTeX-1.40.21 CreationDate: Thu Sep 2 05:40:43 2021 CEST ModDate: Thu Sep 2 05:45:27 2021 CEST Tagged: no UserProperties: no Suspects: no Form: none JavaScript: no Pages: 10 Encrypted: no Page size: 595.276 x 841.89 pts (A4) Page rot: 0 File size: 950039 bytes Optimized: no PDF version: 1.7 name type encoding emb sub uni object ID ------------------------------------ ----------------- ---------------- --- --- --- --------- KNLZHK+VnURWPalladioL Type 1 Custom yes yes no 10 0 KAWSHN+URWPalladioL-Ital Type 1 Custom yes yes no 15 0 LDQNEZ+URWPalladioL-Bold Type 1 Custom yes yes no 20 0 CTQZAY+URWPalladioL-Roma Type 1 Custom yes yes no 24 0 ZPCLNL+CMSY10 Type 1 Builtin yes yes no 28 0 XPUSBJ+TimesNewRomanPS-BoldItalicMT Type 1C WinAnsi yes yes no 36 0 BSDWLC+TimesNewRomanPSMT Type 1C WinAnsi yes yes no 39 0 MCBCCI+SourceSansRoman-Regular Type 1C Custom yes yes no 61 0 MCBCCM+SourceSansRoman.613wght Type 1C Custom yes yes no 65 0 MCBCCO+SourceSansRoman-Bold Type 1C Custom yes yes no 69 0 URYLRZ+URWPalladioL-BoldItal Type 1 Custom yes yes no 87 0 CFEGJO+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 99 0 CFEGKA+PalatinoLinotype TrueType WinAnsi yes yes no 102 0 CFEGKC+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 105 0 CFEGJO+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 115 0 CFEGKA+PalatinoLinotype TrueType WinAnsi yes yes no 118 0 CFEGKC+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 121 0 CFEGJO+PalatinoLinotype,Bold TrueType WinAnsi yes yes no 132 0 CFEGKA+PalatinoLinotype TrueType WinAnsi yes yes no 135 0 CFEGKC+PalatinoLinotype,Italic TrueType WinAnsi yes yes no 138 0 Jhove (Rel. 1.22.1, 2019-04-17) Date: 2023-05-27 03:10:11 CEST RepresentationInformation: 270279.pdf ReportingModule: PDF-hul, Rel. 1.12.1 (2019-04-17) LastModified: 2023-03-31 12:04:55 CEST Size: 950039 Format: PDF Version: 1.7 Status: Well-Formed and valid SignatureMatches: PDF-hul MIMEtype: application/pdf PDFMetadata: Objects: 254 FreeObjects: 1 IncrementalUpdates: 0 DocumentCatalog: PageLayout: SinglePage PageMode: UseNone Outlines: Item: Title: Introduction Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@44c8afef Item: Title: Methods Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@7b69c6ba Item: Title: Results Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@46daef40 Item: Title: Discussion Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@12f41634 Children: Item: Title: Meaning of the Study Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@13c27452 Item: Title: Risk Factors of Post-Ureteroscopic Urosepsis from Previous Published Literature Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@262b2c86 Item: Title: Comparison with Other ML Studies Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@371a67ec Item: Title: Strengths, Limitations and Areas of Future Research Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@5ed828d Item: Title: Conclusions Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@50d0686 Item: Title: References Destination: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject@7a3d45bd Info: Title: A Machine Learning Predictive Model for Post-Ureteroscopy Urosepsis Needing Intensive Care Unit Admission: A Case Control YAU Endourology Study from Nine European Centres Author: Amelia Pietropaolo, Robert M. Geraghty, Rajan Veeratterapillay, Alistair Rogers, Panagiotis Kallidonis, Luca Villa, Luca Boeri, Emanuele Montanari, Gokhan Atis, Esteban Emiliani, Tarik Emre Sener, Feras Al Jaafari, John Fitzpatrick, Matthew Shaw, Chris Harding and Bhaskar K. Somani Subject: Introduction: With the rise in the use of ureteroscopy and laser stone lithotripsy (URSL), a proportionate increase in the risk of post-procedural urosepsis has also been observed. The aims of our paper were to analyse the predictors for severe urosepsis using a machine learning model (ML) in patients that needed intensive care unit (ICU) admission and to make comparisons with a matched cohort. Methods: A retrospective study was conducted across nine high-volume endourology European centres for all patients who underwent URSL and subsequently needed ICU admission for urosepsis (Group A). This was matched by patients with URSL without urosepsis (Group B). Statistical analysis was performed with ‘R statistical software’ using the ‘randomforests’ package. The data were segregated at random into a 70% training set and a 30% test set using the ‘sample’ command. A random forests ML model was then built with n = 300 trees, with the test set used for internal validation. Diagnostic accuracy statistics were generated using the ‘caret’ package. Results: A total of 114 patients were included (57 in each group) with a mean age of 60 16 years and a male:female ratio of 1:1.19. The ML model correctly predicted risk of sepsis in 14/17 (82%) cases (Group A) and predicted those without urosepsis for 12/15 (80%) controls (Group B), whilst overall it also discriminated between the two groups predicting both those with and without sepsis. Our model accuracy was 81.3% (95%, CI: 63.7 92.8%), sensitivity = 0.80, specificity = 0.82 and area under the curve = 0.89. Predictive values most commonly accounting for nodal points in the trees were a large proximal stone location, long stent time, large stone size and long operative time. Conclusion: Urosepsis after endourological procedures remains one of the main reasons for ICU admission. Risk factors for urosepsis are reasonably accurately predicted by our innovative ML model. Focusing on these risk factors can allow one to create predictive strategies to minimise post-operative morbidity. Keywords: kidney stones; urosepsis; ureteroscopy; laser lithotripsy; urolithiasis; nephrolithiasis; kidney calculi; predictor factors Creator: LaTeX with hyperref Producer: pdfTeX-1.40.21 CreationDate: Thu Sep 02 05:40:43 CEST 2021 ModDate: Thu Sep 02 05:45:27 CEST 2021 ID: 0x1d60f40d64c334b40ee25b1edd581bd3, 0x1d60f40d64c334b40ee25b1edd581bd3 Filters: FilterPipeline: FlateDecode Fonts: Type1: Font: BaseFont: LDQNEZ+URWPalladioL-Bold FontSubset: true FirstChar: 3 LastChar: 150 FontDescriptor: FontName: LDQNEZ+URWPalladioL-Bold Flags: Symbolic FontBBox: -152, -301, 1000, 935 FontFile: true EncodingDictionary: Differences: true Font: BaseFont: URYLRZ+URWPalladioL-BoldItal FontSubset: true FirstChar: 110 LastChar: 110 FontDescriptor: FontName: URYLRZ+URWPalladioL-BoldItal Flags: Symbolic FontBBox: -170, -300, 1073, 935 FontFile: true EncodingDictionary: Differences: true Font: BaseFont: CTQZAY+URWPalladioL-Roma FontSubset: true FirstChar: 2 LastChar: 252 FontDescriptor: FontName: CTQZAY+URWPalladioL-Roma Flags: Symbolic FontBBox: -166, -283, 1021, 943 FontFile: true EncodingDictionary: Differences: true Font: BaseFont: KNLZHK+VnURWPalladioL FontSubset: true FirstChar: 189 LastChar: 226 FontDescriptor: FontName: KNLZHK+VnURWPalladioL Flags: Symbolic FontBBox: -166, -283, 1021, 943 FontFile: true EncodingDictionary: Differences: true Font: BaseFont: ZPCLNL+CMSY10 FontSubset: true FirstChar: 6 LastChar: 48 FontDescriptor: FontName: ZPCLNL+CMSY10 Flags: Symbolic FontBBox: -29, -960, 1116, 775 FontFile: true Font: BaseFont: KAWSHN+URWPalladioL-Ital FontSubset: true FirstChar: 44 LastChar: 121 FontDescriptor: FontName: KAWSHN+URWPalladioL-Ital Flags: Symbolic FontBBox: -170, -305, 1010, 941 FontFile: true EncodingDictionary: Differences: true Pages: Page: Sequence: 1 Annotations: Annotation: Subtype: Link Rect: 35, 759, 176, 795 Annotation: Subtype: Link Rect: 515, 759, 560, 795 Annotation: Subtype: Link Rect: 129, 622, 140, 635 Annotation: Subtype: Link Rect: 505, 610, 516, 623 Annotation: Subtype: Link Rect: 477, 597, 488, 610 Annotation: Subtype: Link Rect: 65, 346, 154, 355 Annotation: Subtype: Link Rect: 35, 334, 75, 343 Annotation: Subtype: Link Rect: 35, 162, 97, 184 Annotation: Subtype: Link Rect: 153, 89, 154, 98 Annotation: Subtype: Link Rect: 35, 77, 147, 86 Annotation: Subtype: Link Rect: 131, 17, 268, 28 Annotation: Subtype: Link Rect: 422, 17, 558, 28 Annotation: Subtype: Link Rect: 35, 482, 87, 501 Page: Sequence: 2 Annotations: Annotation: Subtype: Link Rect: 263, 709, 270, 718 ActionDest: 148 Annotation: Subtype: Link Rect: 273, 709, 280, 718 ActionDest: 148 Annotation: Subtype: Link Rect: 199, 683, 206, 692 ActionDest: 148 Annotation: Subtype: Link Rect: 206, 683, 213, 692 ActionDest: 148 Annotation: Subtype: Link Rect: 310, 658, 317, 667 ActionDest: 148 Annotation: Subtype: Link Rect: 318, 658, 324, 667 ActionDest: 148 Annotation: Subtype: Link Rect: 324, 621, 331, 630 ActionDest: 148 Annotation: Subtype: Link Rect: 549, 570, 556, 579 ActionDest: 148 Annotation: Subtype: Link Rect: 223, 533, 230, 542 ActionDest: 148 Annotation: Subtype: Link Rect: 260, 445, 267, 454 ActionDest: 148 Annotation: Subtype: Link Rect: 462, 420, 469, 429 ActionDest: 148 Annotation: Subtype: Link Rect: 258, 369, 265, 379 ActionDest: 148 Annotation: Subtype: Link Rect: 359, 319, 371, 328 ActionDest: 148 Annotation: Subtype: Link Rect: 207, 269, 219, 278 ActionDest: 148 Page: Sequence: 3 Annotations: Annotation: Subtype: Link Rect: 210, 634, 222, 643 ActionDest: 148 Annotation: Subtype: Link Rect: 437, 405, 444, 417 ActionDest: 84 Page: Sequence: 4 Annotations: Annotation: Subtype: Link Rect: 296, 181, 303, 193 ActionDest: 91 Annotation: Subtype: Link Rect: 324, 181, 331, 193 ActionDest: 91 Annotation: Subtype: Link Rect: 438, 131, 445, 143 ActionDest: 126 Annotation: Subtype: Link Rect: 464, 131, 560, 143 Annotation: Subtype: Link Rect: 166, 118, 325, 130 Page: Sequence: 5 Page: Sequence: 6 Annotations: Annotation: Subtype: Link Rect: 298, 163, 310, 172 ActionDest: 148 Annotation: Subtype: Link Rect: 486, 87, 498, 96 ActionDest: 148 Page: Sequence: 7 Annotations: Annotation: Subtype: Link Rect: 234, 685, 246, 694 ActionDest: 148 Annotation: Subtype: Link Rect: 452, 609, 464, 618 ActionDest: 148 Annotation: Subtype: Link Rect: 448, 584, 460, 593 ActionDest: 148 Annotation: Subtype: Link Rect: 411, 534, 423, 543 ActionDest: 148 Annotation: Subtype: Link Rect: 448, 484, 459, 493 ActionDest: 148 Annotation: Subtype: Link Rect: 264, 408, 276, 417 ActionDest: 148 Annotation: Subtype: Link Rect: 312, 333, 324, 342 ActionDest: 148 Annotation: Subtype: Link Rect: 544, 283, 556, 292 ActionDest: 148 Annotation: Subtype: Link Rect: 271, 220, 283, 229 ActionDest: 148 Annotation: Subtype: Link Rect: 273, 182, 285, 191 ActionDest: 148 Annotation: Subtype: Link Rect: 529, 132, 541, 141 ActionDest: 151 Annotation: Subtype: Link Rect: 355, 94, 366, 103 ActionDest: 151 Annotation: Subtype: Link Rect: 359, 69, 371, 78 ActionDest: 151 Page: Sequence: 8 Annotations: Annotation: Subtype: Link Rect: 268, 722, 280, 731 ActionDest: 151 Annotation: Subtype: Link Rect: 298, 534, 310, 543 ActionDest: 151 Page: Sequence: 9 Annotations: Annotation: Subtype: Link Rect: 494, 706, 531, 717 Annotation: Subtype: Link Rect: 145, 673, 182, 683 Annotation: Subtype: Link Rect: 214, 649, 398, 660 Annotation: Subtype: Link Rect: 468, 614, 505, 625 Annotation: Subtype: Link Rect: 511, 614, 547, 625 Annotation: Subtype: Link Rect: 311, 591, 349, 602 Annotation: Subtype: Link Rect: 355, 591, 391, 602 Annotation: Subtype: Link Rect: 131, 569, 169, 579 Annotation: Subtype: Link Rect: 175, 569, 211, 579 Annotation: Subtype: Link Rect: 434, 557, 472, 568 Annotation: Subtype: Link Rect: 95, 523, 133, 533 Annotation: Subtype: Link Rect: 390, 499, 428, 510 Annotation: Subtype: Link Rect: 153, 466, 190, 476 Annotation: Subtype: Link Rect: 520, 430, 557, 441 Annotation: Subtype: Link Rect: 59, 420, 95, 430 Annotation: Subtype: Link Rect: 86, 385, 124, 395 Annotation: Subtype: Link Rect: 95, 351, 133, 361 Annotation: Subtype: Link Rect: 309, 315, 347, 326 Annotation: Subtype: Link Rect: 353, 315, 389, 326 Annotation: Subtype: Link Rect: 283, 269, 320, 280 Annotation: Subtype: Link Rect: 326, 269, 362, 280 Annotation: Subtype: Link Rect: 206, 247, 243, 257 Annotation: Subtype: Link Rect: 250, 247, 286, 257 Annotation: Subtype: Link Rect: 386, 223, 424, 234 Annotation: Subtype: Link Rect: 476, 200, 514, 211 Annotation: Subtype: Link Rect: 280, 177, 317, 188 Annotation: Subtype: Link Rect: 395, 154, 432, 165 Annotation: Subtype: Link Rect: 340, 120, 378, 131 Annotation: Subtype: Link Rect: 240, 85, 278, 96 Page: Sequence: 10 Annotations: Annotation: Subtype: Link Rect: 322, 709, 360, 720 Annotation: Subtype: Link Rect: 140, 676, 178, 686 Annotation: Subtype: Link Rect: 430, 652, 467, 663 Annotation: Subtype: Link Rect: 473, 652, 509, 663 Annotation: Subtype: Link Rect: 198, 629, 236, 640 Annotation: Subtype: Link Rect: 242, 629, 278, 640 Annotation: Subtype: Link Rect: 519, 617, 557, 628 Annotation: Subtype: Link Rect: 59, 607, 95, 617 Checksum: a9a9f45f Type: CRC32 Checksum: f4547c7945f9d69aa248cdb15deb6f50 Type: MD5 Checksum: 38ce9182a69650d7051b571cc3cc906ca80e1696 Type: SHA-1