Palmer, Matthew R.

4541 Background: PROSPER is a phase 3 National Clinical Trials Network study that accrued pts with clinical stage ≥T2 or TanyN+ RCC of any histology planned for radical/ partial nephrectomy. Pts were randomized to perioperative nivo followed by primary tumor resection and 9 cycles of nivo postoperatively or standard surgery followed by observation. To address concerns that core kidney bx might 1) not provide accurate diagnosis and 2) delay surgery due to adverse events (AEs) or scheduling, the study was amended from requiring all patients to undergo core bx prior to treatment for confirmation of RCC, to only require bx in patients randomized to the nivo arm. Herein, we report the accuracy and safety of primary tumor core bxs in predicting final histology and nuclear grade by both site and central pathology review as well as time from enrollment to surgery. We also report AEs of preoperative nivo. Methods: Concordance of both core bx and primary tumor by site and central pathology review of histology and grade (1-2 vs 3-4) are reported, along with the Cohen’s Kappa value, which measures the agreement and concordance (kappa=0 is no concordance and 1 is highest). AEs relating to core bxs and preoperative nivo, as well as time from enrollment to surgery for each arm, comparing pre- and post-amendment (dropping bx requirement in surgery alone arm) are also reported. Results: 387/404 pts in the nivo arm and 171/415 pts in the surgery alone arm had core bxs. 632 patients had both central pathology and site review available. 41 of all randomized patients (819) were considered as non-RCC and 26/41 were identified via bx. The median times from enrollment to surgery for nivo and control arms pre-amendment were 32d vs 19 d, and post-amendment were 21 d vs 14 d, respectively. The median (25th-75th percentile) number of days from last preoperative nivo to surgery was 14 d (9-20). AEs related to core bx, generally from bleeding, were reported in 13/558 (2.3%) pts. 2/13 bxs resulted in life-threatening complications. 21/353 (6%) of pts receiving nivo pre-surgery had ≥ grade 3-5 AE attributed to nivo. 181/353 (51%) pts had any grade AE attributed to nivo. Concordance between bxs and primary tumor pathologies for determining histological subtype was Kappa = 0.62. Agreement between central pathology and originating site review of primary tumor for determining nuclear grade was Kappa = 0.56, and concordance of histology was Kappa = 0.78. Conclusions: The PROSPER trial use of core bxs in advance of neoadjuvant therapy was generally safe, largely consistent with primary tumor histology and grade, and did not delay resection of the primary tumor. AEs of preoperative nivo were consistent with nivo AEs in metastatic disease. This approach is valid for future neoadjuvant trials. Clinical trial information: NCT03055013 .

BackgroundPretreatment determination of renal cell carcinoma aggressiveness may help to guide clinical decision‐making.PurposeTo evaluate the efficacy of residual convolutional neural network using routine MRI in differentiating low‐grade (grade I–II) from high‐grade (grade III–IV) in stage I and II renal cell carcinoma.Study TypeRetrospective.PopulationIn all, 376 patients with 430 renal cell carcinoma lesions from 2008–2019 in a multicenter cohort were acquired. The 353 Fuhrman‐graded renal cell carcinomas were divided into a training, validation, and test set with a 7:2:1 split. The 77 WHO/ISUP graded renal cell carcinomas were used as a separate WHO/ISUP test set.Field Strength/Sequence1.5T and 3.0T/T2‐weighted and T1 contrast‐enhanced sequences.AssessmentThe accuracy, sensitivity, and specificity of the final model were assessed. The receiver operating characteristic (ROC) curve and precision‐recall curve were plotted to measure the performance of the binary classifier. A confusion matrix was drawn to show the true positive, true negative, false positive, and false negative of the model.Statistical TestsMann–Whitney U‐test for continuous data and the chi‐square test or Fisher's exact test for categorical data were used to compare the difference of clinicopathologic characteristics between the low‐ and high‐grade groups. The adjusted Wald method was used to calculate the 95% confidence interval (CI) of accuracy, sensitivity, and specificity.ResultsThe final deep‐learning model achieved a test accuracy of 0.88 (95% CI: 0.73–0.96), sensitivity of 0.89 (95% CI: 0.74–0.96), and specificity of 0.88 (95% CI: 0.73–0.96) in the Fuhrman test set and a test accuracy of 0.83 (95% CI: 0.73–0.90), sensitivity of 0.92 (95% CI: 0.84–0.97), and specificity of 0.78 (95% CI: 0.68–0.86) in the WHO/ISUP test set.Data ConclusionDeep learning can noninvasively predict the histological grade of stage I and II renal cell carcinoma using conventional MRI in a multiinstitutional dataset with high accuracy.Level of Evidence3Technical Efficacy Stage2