miR-371a-3p Predicting Viable Tumor in Patients Undergoing Retroperitoneal Lymph Node Dissection for Metastatic Testicular Cancer: The SWENOTECA-MIR Study

Open AccessJournal of UrologyOriginal Clinical Article25 Jul 2024miR-371a-3p Predicting Viable Tumor in Patients Undergoing Retroperitoneal Lymph Node Dissection for Metastatic Testicular Cancer: The SWENOTECA-MIR StudyThis article is commented on by the following:Editorial Comment Anna Thor, Mette Pernille Myklebust, Anna Grenabo Bergdahl, Per-Olof Lundgren, Viktor Skokic, Bjarte Almås, Hege Sagstuen Haugnes, Torgrim Tandstad, Olof Akre, Gabriella Cohn-Cedermark, Olav Dahl, and Anders Kjellman Anna ThorAnna Thor Corresponding Author: Anna Thor, MD, Division of Urology, Department of Clinical Science, Intervention and Technology, Karolinska Institute, 171 77Stockholm, Sweden (email protected) https://orcid.org/0000-0003-2915-1761 , Mette Pernille MyklebustMette Pernille Myklebust https://orcid.org/0000-0002-3552-3747 Mohn Cancer Research Laboratory, Department of Oncology and Medical Physics, Haukeland University Hospital, Bergen, Norway , Anna Grenabo BergdahlAnna Grenabo Bergdahl https://orcid.org/0000-0003-3266-7611 Department of Urology, Gothenburg University, Gothenburg, Sweden Department of Urology, Sahlgrenska University Hospital, Gothenburg, Sweden , Per-Olof LundgrenPer-Olof Lundgren https://orcid.org/0000-0001-7976-6962 Division of Urology, Department of Clinical Science, Intervention and Technology, Karolinska Institute, Stockholm, Sweden Department of Urology, Pelvic Cancer, Karolinska University Hospital, Stockholm, Sweden , Viktor SkokicViktor Skokic Department of Urology, Pelvic Cancer, Karolinska University Hospital, Stockholm, Sweden Department of Molecular Medicine and Surgery, Karolinska Institute, Stockholm, Sweden , Bjarte AlmåsBjarte Almås https://orcid.org/0000-0003-4484-040X Department of Urology, Haukeland University Hospital, Bergen, Norway , Hege Sagstuen HaugnesHege Sagstuen Haugnes https://orcid.org/0000-0002-9813-0561 Department of Oncology, University Hospital of North Norway, Tromsø, Norway Department of Clinical Medicine, UIT–The Arctic University of Norway, Tromsø, Norway , Torgrim TandstadTorgrim Tandstad https://orcid.org/0000-0002-9786-2138 The Cancer Clinic, St Olavs University Hospital, Trondheim, Norway Department of Clinical and Molecular Medicine, Norwegian University of Science and Technology, Trondheim, Norway , Olof AkreOlof Akre https://orcid.org/0000-0002-3750-3029 Department of Urology, Pelvic Cancer, Karolinska University Hospital, Stockholm, Sweden Department of Molecular Medicine and Surgery, Karolinska Institute, Stockholm, Sweden , Gabriella Cohn-CedermarkGabriella Cohn-Cedermark https://orcid.org/0000-0002-5681-5829 Department of Oncology-Pathology, Karolinska Institute, Stockholm, Sweden Department of Pelvic Cancer, Genitourinary Oncology Unit, Karolinska University Hospital, Stockholm, Sweden , Olav DahlOlav Dahl Department of Oncology, Haukeland University Hospital, Bergen, Norway Department of Clinical Science, Medical Faculty, University of Bergen, Norway , and Anders KjellmanAnders Kjellman https://orcid.org/0000-0002-9080-9027 Division of Urology, Department of Clinical Science, Intervention and Technology, Karolinska Institute, Stockholm, Sweden Department of Urology, Pelvic Cancer, Karolinska University Hospital, Stockholm, Sweden on behalf of the SWENOTECA Working Group View All Author Informationhttps://doi.org/10.1097/JU.0000000000004164AboutAbstractPDF Cite Export CitationSelect Citation formatNLMAMAIEEEACMAPAChicagoMLAHarvardTips on citation downloadDownload citationCopy citation ToolsAdd to favoritesTrack Citations ShareFacebookTwitterLinked InEmail Abstract Purpose: The SWENOTECA-MIR prospective multicenter study aims to assess the clinical value of miR-371a-3p as a novel marker in metastatic germ cell tumor patients undergoing retroperitoneal lymph node dissection (RPLND), to predict the presence of viable residual tumor. Materials and Methods: A total of 114 patients (86 nonseminomas, 28 seminomas) who underwent surgery for presumed metastatic disease pre chemotherapy (primary RPLND) and post chemotherapy RPLND were included. The expression of miR-371a-3p was evaluated using reverse transcription–digital droplet polymerase chain reaction before and after RPLND. Pre- and postoperative miR-371a-3p levels were statistically compared, and optimism-corrected performance calculations compared with conventional serum tumor markers. Associations were evaluated by logistic regression. Patients who underwent primary RPLND were categorized into seminoma and nonseminoma groups. Results: Among the seminoma patients (n = 24) undergoing primary RPLND, all had normal conventional markers. Six patients received adjuvant treatment before surgery. miR-371a-3p exhibited a sensitivity of 74%, specificity of 100%, positive predictive value of 100%, and negative predictive value of 21% for viable tumor. The levels of miR-371a-3p significantly decreased after surgery. In the nonseminoma group (n = 18) treated with primary RPLND, 22% had elevated conventional markers and 3 had received prior adjuvant treatment. miR-371a-3p showed a sensitivity of 34%, specificity of 88%, positive predictive value of 67%, and negative predictive value of 62% for the primary nonseminoma patients. No association was observed between stage or prior adjuvant treatment and the outcome of the miR test. In the postchemotherapy group (n = 72), the miR-371a-3p sensitivity was 9%, reducing to 0 when excluding patients with seminoma (n = 4). Teratomas and benign histology were essentially negative. Conclusions: Our study highlights miR-371a-3p as a fairly sensitive and highly specific marker for prechemotherapy seminomas, outperforming conventional markers. However, in prechemotherapy nonseminomas as well as in postchemotherapy patients, we observed low sensitivity and no significant differences in miR-371a-3p levels before and after surgery, suggesting limited utility for miR-371a-3p in this context. Testicular germ cell tumor (TGCT) is the most common cancer in young males, with rising incidence.1 Although treatment for metastatic TGCT is highly effective, the chemotherapy regimens used are associated with significant long-term side effects, including cardiovascular disease and secondary malignancies.2 Survivors may also experience excess mortality due to prior therapy.3 Staging and monitoring of TGCTs involve repeated CT and MRI, and measuring serum protein biomarkers beta human chorionic gonadotropin (β-hCG), alpha fetoprotein (AFP), and lactate dehydrogenase. However, the markers lack sensitivity and specificity, with 40% of all TGCTs being marker negative, especially seminomas and teratomas.4 Chemotherapy typically includes bleomycin, etoposide, and cisplatin (BEP) courses varying by clinical stage (CS) and risk group. Adjuvant treatment by one course of carboplatin or BEP is offered to selected CS I patients, according to the SWENOTECA protocol. BEP was administered as adjuvant treatment to seminoma patients within the randomized ABC-study.5 Surgery is indicated for nonseminoma patients with a visible residual tumor ≥ 10 mm in the retroperitoneum after first-line chemotherapy,5 while for seminoma patients with small retroperitoneal metastases, surgery is being evaluated as a primary treatment in prospective clinical trials.6-8 Retroperitoneal lymph node dissection (RPLND) is a complex surgical procedure, associated with perioperative morbidity and risk of long-term sequelae, including loss of antegrade ejaculation. In previous studies, 44% to 72% of RPLND patients harbored only necrosis or fibrosis in residual masses.9-11 Reliable diagnostic tools are needed to distinguish patients requiring RPLND from those who can be spared. In 2011, the microRNA clusters miR-302/367 and miR-371 to 373 were identified as potential novel markers for TGCTs.12 Subsequent research revealed miR-371a-3p as the most promising micro-RNA marker, as it has been found to be present in nearly all cases of TGCTs other than teratoma.13-16 Dieckmann et al demonstrated that miR-371a-3p expressed sensitivity and specificity over 90% in TGCTs at various CSs.14 Further studies reported similar results for miR-371a-3p in postchemotherapy patients with retroperitoneal residual masses or relapse.17,18 Teratomas have consistently shown negative results in miR-371a-3p measurements across various studies.19 To determine the clinical value of the lead candidate miR-371a-3p as a tumor marker, large prospective clinical trials are required. The first part of the SWENOTECA-MIR study on miR-371a-3p in 180 orchiectomy patients showed better performance than conventional markers, with an overall sensitivity of 89%.20 This study aims to assess miR-371a-3p expression before and after RPLND in TGCT patients, to evaluate its accuracy and potential as a predictor of viable disease, and to see whether cytoreductive surgery affects its values. METHODS Study Design and Participants The SWENOTECA-MIR trial is a prospective binational multicenter study, with 3 cohorts based on different interventions: orchiectomy, chemotherapy, or RPLND. The primary outcomes are miR-371a-3p levels before and after each intervention. Inclusion criteria were males aged 18 to 70 without prior malignancy, diagnosed with TGCT, and planned for RPLND, either due to presumed metastatic disease pre chemotherapy (primary RPLND), or postchemotherapy RPLND. Exclusion criteria included previous malignancy or inability to understand the consent form due to a language barrier. This cohort comprises 114 patients who underwent RPLND from March 2017 to October 2022 in 3 tertiary hospitals. Unilateral or bilateral, open or robot-assisted RPLND was performed in templates according to SWENOTECA guidelines5 by a small group of experienced surgeons affiliated with the study. All oncologists and urologists involved in patient treatment were kept blinded to the results. Clinical parameters including levels of AFP, β-hCG, and lactate dehydrogenase, CS according to Royal Marsden, age at RPLND, radiology examinations, orchiectomy and RPLND histology, and treatments were retrieved from medical records. The study was approved by the Regional Ethics Committees (REC Stockholm 2018/1730-31 and REC Central Norway 2015/1475). All patients received oral and written information and gave written consent. Laboratory Methods Study samples were collected up to 1 week prior to and 18 to 24 hours after RPLND. Serum was stored at −80 °C until analysis. RNA was extracted from 200 µL serum using the miRNeasy Kit (Qiagen, P/N 217004). Reverse transcription (RT) was performed using the specific RT primers from the TaqMan assays miR-371a-3p (Thermo Fisher Scientific, ID 002124) and miR-30b-5p (Thermo Fisher Scientific, ID 000602). RT-digital droplet (dd) polymerase chain reaction (PCR) was performed using a QX200 AutoDG Droplet Digital PCR System for 96 well plates (Bio-Rad) and Supermix for Probes (Bio-Rad, P/N 186-024), as described previously.20 The endogenously expressed miR-30b-5p was used as an internal control. Results for microRNA-371a-3p are given as copies per µL serum. Details regarding the analysis can be found in the Supplementary Protocol. The threshold for defining a sample as positive for miR-371a-3p was 0.45 copies per μL serum, as previously described.20 Statistical Analysis Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for the preoperative sample of miR-371a-3p and the conventional markers β-hCG/AFP. The gold standard in our calculations of sensitivity, specificity, PPV, and NPV is the postoperative pathology report of prevalence of viable tumor in the retroperitoneal lymph nodes. Teratomas were anticipated to yield true negative results in miR-371a-3p tests and were categorized as benign outcomes or disease-negative results in the performance calculations. To assess the performance of our predictive model and account for potential overfitting, bootstrap optimism correction was applied. To that end, the analysis was reformulated in terms of classification, treating preoperative positive miR-371a-3p and preoperative positive conventional tumor markers as predictors, and viable cancer in the pathohistological report as the outcome. A majority class classifier was used. In cases of equal estimated outcome class probabilities when training on a bootstrap sample, classes were randomly assigned with probability 0.5. In the cases of absence of a class in the predictor in a bootstrap sample, the overall class probabilities in the original outcome sample were used when applying the classifier to the original predictor values. A total of 10,000 bootstrap samples were used in the performance calculations. Paired violin plots for histological outcomes were constructed according to miR-371a-3p levels before and after surgery, for the subgroups seminomas, nonseminomas, and benign/teratomas. Statistical significance between pre- and postoperative miR-371a-3p levels was determined using the Wilcoxon signed-rank test due to nonnormal distribution of miR-371a-3p values. For the primary RPLND patients, logistic regression was used to assess associations between a high tumor stage or the absence of adjuvant chemotherapy, with a preoperative positive miR-371a-3p test. Bootstrapping of 1000 replications was employed to obtain robust estimates of the associations. Statistical significance was achieved at a level of P < .05 for 2-tailed analyses. Data analysis and graphics were created using STATA version 16.1 (StataCorp LCC, College Station, Texas) and R version 4.3.2 (The R Foundation, Vienna, Austria). RESULTS In total, the study included 114 patients, of whom 86 had nonseminomas and the remaining 28 had seminomas. The cohort was categorized into 2 groups based on the clinical indication for RPLND, primary (n = 42) or post chemotherapy (n = 72). Among those who underwent primary RPLND for early-stage disease, further subgroups were created for seminomas (n = 24) and nonseminomas (n = 18; Figure 1). Figure 1. Comprehensive overview of the study cohort's characteristics and outcomes by histology and tumor markers. AFP indicates alpha fetoprotein; GCT, germ cell tumor; hCG, human chorionic gonadotropin; pos, positive; POST-CHEMO, postchemotherapy; RPLND, retroperitoneal lymph node dissection. Download PPTIn the 18 patients with CS I or II nonseminoma treated with RPLND, postoperative histopathology revealed viable cancers other than teratoma in 8 (44%; Table 1). Among these, 4 (22%) expressed positive conventional markers at surgery, while only 3 patients tested positive for miR-371a-3p. Two patients with viable nonseminoma were positive for both conventional markers and miR-371a-3p at surgery. However, one patient with teratoma showed a postoperative miR-371a-3p positive result after testing negative preoperatively, indicating a false-positive result. No significant difference was observed in levels of miR-371a-3p pre- or postoperatively, within this group (P = .6) (Figure 2). Logistic regression was deemed inappropriate because of the low count of positive tests. The miR-371a-3p test had an optimism-corrected sensitivity of 34%, specificity of 88%, PPV of 67%, and NPV of 62%. The conventional markers had a sensitivity of 49%, specificity of 100%, PPV 98%, and NPV 63%. Among the 3 patients with a viable cancer and a positive miR-371-3p test result, one patient had received one course of BEP before surgery, while the other 2 were chemotherapy naïve. Table 1. Patients With Nonseminoma Treated With Retroperitoneal Lymph Node Dissection for Presumed Metastatic Germ Cell Tumor (Nonseminoma Primary Retroperitoneal Lymph Node Dissection Group, n = 18) Clinical stage, No. (%) CS I with somatically differentiated GCT 3 (17) CS IIA 4 (22) CS IIB 9 (50) CS IIC 1 (5.5) Unknown 1 (5.5) Orchiectomy specimen histology, No. (%) Embryonal cancer 4 (22) Teratoma 4 (22) Mixed 8 (44) With teratoma 5 With seminoma 4 With somatical differentiation 4 Unspecified nonseminoma tumor 2 (11) Chemotherapy, No. (%) Adjuvant (BEP × 1) 3 (17) No chemotherapy 15 (83) Serum tumor markers positive at RPLND, No. (%) 4 (22) AFP elevated 1 AFP and β-hCG elevated 1 Unknown 2 Marker negative 14 (78) Age at RPLND, median (range), y 29 (18-61) RPLND specimen histology, No. (%) Teratoma 7 (39) Benign 3 (17) Viable tumor 8 (44) Embryonal carcinoma 3 Seminoma 1 Mixed tumor embryonal carcinoma + yolk sac 2 Somatically differentiated GCT 1 Yolk sac tumor 1 Sensitivity serum tumor markers, optimism-corrected (%) 49 Specificity serum tumor markers, optimism-corrected (%) 100 PPV serum tumor markers, optimism-corrected (%) 98 NPV serum tumor markers, optimism-corrected (%) 63 True positive miR-371a-3p, No. (%) 3 (38) False-positive miR-371a-3p, No. (%) 1 (13) False-negative miR-371a-3p, No. (%) 4 (50) Sensitivity miR-371a-3p optimism-corrected (%) 34 Specificity miR-371a-3p, optimism-corrected (%) 88 PPV miR-371a-3p, optimism-corrected (%) 67 NPV miR-371a-3p, optimism-corrected (%) 62 Abbreviations: AFP, alpha fetoprotein; β-hCG, beta human chorionic gonadotropin; BEP, bleomycin, etoposide, cisplatin; CS, clinical stage; GCT, germ cell tumor; NPV, negative predictive value; PPV, positive predictive value; RPLND, retroperitoneal lymph node dissection. Figure 2. Violin plot illustrating the pre- and postoperative levels of miR-371a3p in patients treated with primary retroperitoneal lymph node dissection (RPLND), categorized by histological outcomes. A positive (pos) threshold is defined as miR-371a-3p exceeding 0.45 copies/µL serum. ddPCR indicates digital droplet polymerase chain reaction. Download PPTThe primary seminoma group included 24 patients (Table 2). Among them, 42% (10 patients) were relapses in CS I, with 6 receiving adjuvant carboplatin or BEP prior to primary RPLND. All were negative in conventional markers at surgery. Postoperative histology identified viable cancers in 23, and 1 benign case. Among the viable seminomas 17 tested positive for miR-371a-3p. The optimism-corrected sensitivity was 74%, specificity 100%, PPV 100%, and NPV 21%. Logistic regression analysis indicated no association between a higher CS and a positive miR-371a-3p test (odds ratio OR 0.7 95% CI: 0.1-9, P = .8) or prior adjuvant treatment and a positive miR-371a-3p test (OR 2.1 95% CI: 0.2-23, P = .6). After adding bootstrapping to the logistic regression, the results remained virtually unchanged, both for CS (OR 0.7 95% CI: 0.1-4, P = .7) and prior adjuvant treatment (OR 2.1 95% CI: 0.4-12, P = .4). The preoperative sample measurements of miR-371a-3p in patients with viable seminomas were significantly higher than the postoperative, showing a rapid decline in serum levels after surgery (P = .001; Figures 2 and 3). Details of patients with postoperative viable disease, excluding teratomas, are displayed in Tables 3 and 4. Table 2. Patients With Seminoma Treated With Retroperitoneal Lymph Node Dissection for Presumed Metastatic Germ Cell Tumor (Seminoma Primary Retroperitoneal Lymph Node Dissection Group, n = 24) Clinical stage, No. (%) CS I with progression 10 (42) CS IIA 12 (50) CS IIB 2 (8) Chemotherapy, No. (%) Adjuvant 6 (25) Carboplatin × 1 5 BEP × 1 1 No chemotherapy 18 (75) Serum tumor markers positive at RPLND, No. (%) 0 (0) Median age at RPLND, median (range), y 42 (29-53) RPLND specimen histology, No. (%) Benign 1 (4) Seminoma 23 (96) True positive miR-371a-3p, No. (%) 17 (74) False-negative miR-371a-3p, No. (%) 6 (26) Sensitivity miR-371a-3p, optimism-corrected (%) 74 Specificity miR-371a-3p, optimism-corrected (%) 100 PPV miR-371a-3p, optimism-corrected (%) 100 NPV miR-371a-3p, optimism-corrected (%) 21 Abbreviations: BEP, bleomycin, etoposide, cisplatin; CS, clinical stage; GCT, germ cell tumor; NPV, negative predictive value; PPV, positive predictive value; RPLND, retroperitoneal lymph node dissection. Figure 3. miR-371-3p levels pre and post retroperitoneal lymph node dissection (RPLND) for patients with outcomes of viable seminoma. The decrease in miR-371a-3p is significant (P = .001). Red indicates patients having received prior chemotherapy. ddPCR indicates digital droplet polymerase chain reaction. Download PPT Table 3. Pt No. Group Clinical stage at diagnosis Orchiectomy histology Testicular tumor size (mm) Prior chemotherapy Age at RPLND (y) Clinical stage at RPLND RPLND histology No. positive lymph nodes Largest tumor diameter (mm) Positive markers at surgery Pre miR-371a-3p copies/µL serum (threshold 0.45) Post miR-371a-3p copies/µL serum (threshold 0.45) Positive miR-371a-3p test 1 NS IIB EC - 0 48 IIB YST 1 10.0 Yes 0.00 0.00 No 2 NS IVC - - PEI × 4 48 IVC YST, T 2 05.5 No 0.00 0.18 No 3 NS IIC YST 15.0 PEI × 4 25 IIB YST, T 2 05.0 No 0.00 0.00 No 4 NS IIC S (AFP+) 15.0 BEP × 3 30 IIC YST, T 3 11.0 No 0.00 0.00 No 5 NS IIB EC, YST, T, S 47.0 0 41 IIB S 1 27.0 Yes 0.12 0.20 No 6 NS IIA EC, YST, S - 0 24 IIA EC 1 03.2 No 0.06 0.00 No 7 NS IVC - - PEI × 2, TIP × 2, HD × 2 24 IVC CC - - Yes 0.25 0.29 No 8 NS I - 16.0 BEP × 1 51 IIA EC, YST, undefined malignant GCT - - Yes 103.7 19.9 Yes 9 NS IIA - 40.0 0 51 IIA EC, YST - - Yes 38.0 5.70 Yes 10 NS IIC YST, T 45.0 BEP × 4 20 IIC YST, T - - Yes 0.21 0.07 No 11 NS IIA EC 11.0 0 21 IIA EC - - No 0.43 0.21 No 12 NS IIA T 08.0 0 18 IIB EC - - No 9.51 0.21 Yes 13 NS IIB EC, YST, T 40.0 BEP × 3 30 IIB YST - - No 0.00 0.15 No 14 NS I T, SOMATIC 29.0 0 49 I SOMATIC - - No 0.12 0.19 No Abbreviations: AFP+, alpha fetoprotein serum marker positive; BEP, bleomycin, etoposid, cisplatin; EC, embryonal carcinoma; GCT germ cell tumor; HD, high-dose chemotherapy; NS, nonseminoma; PEI, cisplatin, etoposid, ifosfamide; Pt, patient; RPLND, retroperitoneal lymph node dissection; S, seminoma; SOMATIC, somatically diffentiated germ cell tumor; T, teratoma; TIP, paclitaxel, ifosfamid, cisplatin; YST, yolk sac tumor. Characteristics of nonseminoma patients with viable histology results other than teratoma after RPLND (n = 14). Table 4. Pt No. Group Clinical stage at diagnosis Orchiectomy histology Testicular tumor size (mm) Prior chemotherapy Age at RPLND (y) Clinical stage at RPLND RPLND histology No. positive lymphnodes Largest tumor diameter (mm) Positive markers at surgery Pre-op miR-371a-3p copies/µL serum (threshold 0.45) Post-op miR-371a-3p copies/µL serum (threshold 0.45) Positive miR-371a-3p test 1 S I S 24.0 Carbo × 1 41 IIA S 1 14.0 No 0.63 0.21 Yes 2 S I S 19.0 0 46 IIA S 4 16.0 No 0.51 0.12 Yes 3 S IIA S 58.0 0 42 IIA S 1 11.5 No 0.63 0.00 Yes 4 S I S 12.0 0 41 IIA S 1 15.0 No 0.00 0.06 No 5 S IIA Burned out - 0 29 IIA S 3 08.0 No 0.39 0.00 No 6 S I S 22.0 Carbo × 1 46 IIA S 1 - No 1.71 0.12 Yes 7 S IIA S 82.0 0 30 IIA S 2 20.0 No 2.97 0.33 Yes 8 S I S 52.0 0 48 IIB S 2 24.0 No 4.14 14.9 Yes 9 S I S 90.0 Carbo × 1 51 IIB S 1 13.0 No 2.31 0.12 Yes 10 S IIA S 60.0 BEP × 1 47 IIA S 1 19.0 No 1.38 0.24 Yes 11 S IIA S 80.0 0 36 IIA S 1 02.5 No 0.12 0.12 No 12 S IIA S 25.0 0 52 IIA S 1 19.0 No 0.84 0.18 Yes 13 S IIA S 30.0 0 39 IIA S 1 19.0 No 0.18 0.06 No 14 S I S 35.0 0 32 IIB S 1 10.7 No 2.99 1.53 Yes 15 S I S - 0 34 IIA S 1 20.0 No 0.57 0.40 Yes 16 S I S - Carbo × 1 32 IIC S 1 25.0 No 0.06 0.00 No 17 S I S 70.0 0 38 IIA S - - No 0.47 0.54 Yes 18 S I S 52.0 Carbo × 1 53 IIA S - - No 0.72 0.18 Yes 19 S I S 45.0 0 35 IIA S - - No 1.50 0.06 Yes 20 S I S 48.0 0 47 IIA S - - No 1.35 0.18 Yes 21 S IIA S 38.0 0 43 IIA S - - No 0.06 0.12 No 22 S I S 32.0 0 51 IIA S - - No 0.78 - Yes 23 S IIA S 80.0 0 45 IIA S - - No 0.71 0.12 Yes Abbreviations: BEP, bleomycin, etoposide, cisplatin; Carbo, carboplatin; Pt, patient; RPLND, retroperitoneal lymph node dissection; S, seminoma. Characteristics of seminoma patients with viable histology results after primary RPLND (n = 23). In the postchemotherapy group of 72 patients (68 nonseminomas, 4 seminomas) 65% were CS II at diagnosis (Table 5). Indications for postchemotherapy RPLND in seminoma patients were progression of residual tumor, inconclusive radiology, or elevated AFP levels suggesting nonseminomatous elements. Overall, 15% had elevated conventional markers at surgery. The postoperative histology revealed 8 (11%) viable cancers, 38 (53%) teratomas, and 26 (36%) benign findings. Among the 8 viable cancers, 2 expressed conventional markers. One patient had a positive result in miR-371a-3p. The histology revealed a large seminoma tumor of 5 cm, and the patient was negative in conventional markers. Table 5. Patients Treated With Postchemotherapy Retroperitoneal Lymph Node Dissection (Postchemotherapy Group, n = 72) Orchiectomy specimen histology, No. (%) Nonseminoma 68 (94) Seminoma 4 (6) Clinical stage, No. (%) CS II A-C 47 (65) CS III A-C 7 (10) CS IV A-C 18 (25) Chemotherapy, No. (%) Standard (BEP, PEI, EP; × 3-4) 61 (85) Intensified (TIP, GOP, EMA/CO) 10 (14) Unknown 1 (1) Serum tumor markers positive at RPLND, No. (%) 11 (15) AFP elevated 4 β-hCG elevated 2 AFP and β-hCG elevated 1 Elevated serum tumor markers, unknown type 4 Unknown 1 (1) Marker negative 60 (83) Age at RPLND, years median (min-max) 29 (18-58) RPLND specimen histology, No. (%) Teratoma 38 (53) Benign 26 (36) Viable tumor 8 (11) Yolk sac tumor 1 Choriocarcinoma 1 Seminoma 2 Mixed tumor teratoma + yolk sac tumor 4 Sensitivity serum tumor markers, optimism-corrected (%) 36 Specificity serum tumor markers, optimism-corrected (%) 87 PPV serum tumor markers, optimism-corrected (%) 0 NPV serum tumor markers, optimism-corrected (%) 67 True positive miR-371a-3p, No. (%) 1 (13) False-negative miR-371a-3p, No. (%) 7 (88) Sensitivity miR-371a-3p, optimism-corrected (%) 9 Specificity miR-371a-3p, optimism-corrected (%) 100 PPV miR-371a-3p, optimism-corrected (%) 100 NPV miR-371a-3p, optimism-corrected (%) 64 Abbreviations: AFP, alpha fetoprotein; β-hCG, beta human chorionic gonadotropin; CS, clinical stage; EMA/CO, etoposide, methotrexate, actinomycin D, cyclophosphamide, vincristine; EP, cisplatin, etoposide; GOP, gemcitabine, oxaliplatin, paclitaxel; max, maximum; min, minimum; NPV, negative predictive value; PEI, cisplatin, etoposide, ifosfamide; PPV, positive predictive value; RPLND, retroperitoneal lymph node dissection; TIP, paclitaxel, ifosfamide, cisplatin. The optimism-corrected diagnostic performance calculations of miR-371a-3p for the postchemotherapy group yielded sensitivity 9%, specificity 100%, PPV 100%, and NPV 64%. The sensitivity was reduced to 0 when the seminoma patients were excluded. The conventional markers exhibited sensitivity of 36%, specificity of 87%, PPV of 0%, and NPV of 67% within the entire group of postchemotherapy patients. Table 6 presents the performance metrics calculated using both standard methods and bootstrap resampling. Table 6. Sensitivity Specificity PPV NPV Primary nonseminoma group (n = 18) miR371a-3p Original estimates 0.38 0.9 0.75 0.64 Optimisms 0.04 0.02 0.08 0.03 Optimism-corrected estimates 0.34 0.88 0.67 0.62 Conventional tumor markers Original estimates 0.5 1 1 0.69 Optimisms 0.01 0.00 0.02 0.07 Optimism-corrected estimates 0.49 1 0.98 0.63 Primary seminoma group (n = 24) miR371a-3p Original estimates 0.74 1 1 0.14 Optimisms 0.00 0.00 0.00 −0.07 Optimism-corrected estimates 0.74 1 1 0.21 Conventional tumor markers Original estimates 0 1 - 0.04 Optimisms 0 0 −6.7 - Optimism-corrected estimates 0 1 - - Postchemotherapy group (n = 72) miR371a-3p Original estimates 0.13 1 1 0.9 Optimisms 0.04 0 0 0.26 Optimism-corrected estimates 0.09 1 1 0.64 Conventional tumor markers Original estimates 0.38 0.87 0.27 0.92 Optimisms 0.01 0.00 0.28 0.25 Optimism-corrected estimates 0.36 0.87 −0.01 0.67 Abbreviations: NPV, negative predictive value; PPV, positive predictive value. Optimism-corrected estimates for performance calculations using bootstrap resampling. DISCUSSION This study represents the largest evaluation to date on the utility of miR-371a-3p predicting histological outcomes post RPLND. Given centralized treat