Ho, Thai H.

694 Background: The treatment landscape of localized renal cell carcinoma (RCC) is rapidly evolving. Recent trials have demonstrated contrasting efficacy with adjuvant immunotherapy. Therefore, we sought to assess the mixed treatment comparisons among different adjuvant treatment options using updated data from trials and quantified the absolute effect with these treatments stratified by risk categories. Methods: This living network meta-analysis was conducted using the living interactive evidence (LIvE) synthesis framework. We used stratified baseline risks of disease progression from observational data and at 5, 10, 15 years from the Leibovich risk stratification system (based on risk scores ranging from 0 to ≥15). Corresponding intervention risks were then approximated using relative effect estimates (from NMA) and baseline risks. The difference between CIRs and baseline risks were calculated to present absolute risk differences in each risk category. Results: This NMA included eight RCTs with 8480 participants and seven unique treatment options. Pembrolizumab (pembro; rank 1) was associated with improved disease-free survival (DFS) when compared to atezolizumab (atezo; rank 6; hazard ratio: 0.68; 0.49;0.93), and nivolumab-ipilimumab (nivoipi; rank 5; 0.68; 0.48-0.97). However, no statistically significant difference was observed between pembro and atezo for overall survival (0.53; 0.28-1.01). Survival data for nivoIpi was not reported. No new statistically significant differences were observed since last update. The absolute benefit of atezo and nivoipi was minimal with higher T and N patients (Table). The results were similar with increasing Leibovich risk scores. Conclusions: Current evidence favors the use of a risk adapted approach when offering adjuvant immunotherapy. Adjuvant pembrolizumab remains a preferred treatment in patients with RCC who underwent nephrectomy. [Table: see text]

692 Background: Sarcomatoid mRCC exhibits poor prognosis and limited response to vascular endothelial growth factor pathway inhibition. Therefore, we assessed the efficacy of ICI combination therapy in this patient population using data from contemporary trials. Methods: MEDLINE and EMBASE were searched to identify phase III randomized controlled trials (RCTs) comparing the efficacy of ICI combinations with sunitinib monotherapy in patients with sarcomatoid mRCC. Patient important outcomes of interest included overall survival (OS), progression-free survival (PFS), objective response rate (ORR), and complete response (CR). Precomputed hazard ratios (HR) with 95% confidence intervals (CI) for survival outcomes and binary outcome data for response rates (expressed as relative risk [RR] were meta-analyzed using a DerSimonian-Lairds random-effects method. Mixed treatment comparisons among different ICI combinations were made using a network-meta-analysis within the Bayesian framework. The surface under the cumulative ranking curves (SUCRA) were computed to assess the relative treatment rankings. Results: Six RCTs with a total of 618 patients were considered eligible for inclusion. ICI combination therapy was significantly associated with improved OS (HR: 0.56; 95% CI: 0.43-0.72) and PFS (HR: 0.50; 0.40-0.62), increased ORR (RR: 2.42; 1.92-3.06), and CR (RR: 4.23; 2.00-8.93) when compared to sunitinib in patients with sarcomatoid mRCC (Table). The results were consistent with Hartung-Knapp adjustment. Mixed treatment comparisons using current data revealed no statistically significant differences among different ICI combinations. However, the combination of nivolumab-ipilimumab was consistently ranked higher (rank 2 for OS, ORR, and CR) and may potentially be more efficacious than other counterparts. Conclusions: Current evidence suggests improved survival, delayed disease progression and increased response rates with the use of ICI combination therapy in patients with sarcomatoid mRCC. [Table: see text]

Accurate clinical staging of bladder cancer aids in optimizing the process of clinical decision-making, thereby tailoring the effective treatment and management of patients. While several radiomics approaches have been developed to facilitate the process of clinical diagnosis and staging of bladder cancer using grayscale computed tomography (CT) scans, the performances of these models have been low, with little validation and no clear consensus on specific imaging signatures. We propose a hybrid framework comprising pre-trained deep neural networks for feature extraction, in combination with statistical machine learning techniques for classification, which is capable of performing the following classification tasks: (1) bladder cancer tissue vs. normal tissue, (2) muscle-invasive bladder cancer (MIBC) vs. non-muscle-invasive bladder cancer (NMIBC), and (3) post-treatment changes (PTC) vs. MIBC.

e16550 Background: Improved computational power and modern algorithms have generated significant interest in radiomics for cancer diagnosis and staging. Here we assess the performance of deep learning (DL) models as a means for feature extraction in combination with supervised machine learning (ML) algorithms for accurate staging and chemotherapy response assessment of bladder cancer. Methods: Deidentified grayscale CT images from bladder cancer patients scheduled to undergo radical cystectomy were included in this retrospective study. These images were manually annotated with two regional masks (normal region and cancer region). Five DL models- namely, AlexNet, GoogleNet, InceptionV3, ResNet-50, and XceptionNet pre-trained on the ImageNet dataset, a public dataset, were then fine-tuned on our bladder CT scan data to extract features. Through feature selection process, the subset of the features was used to build ML classifiers for classification. The classification was performed using five different ML classifiers, namely k-Nearest Neighbor (KNN), Naïve-Bayes (NB), Support Vector Machine (SVM), Linear Discriminant Analysis (LDA), and Decision Tree (DT). The classification task was performed with 10-fold cross-validation, and each of the experiments contained a different but not mutually exclusive subset of samples. The evaluation metrics include accuracy, sensitivity, specificity, precision, and F1-score. Results: A total of 200 deidentified grayscale CT images of 100 patients with histologically proven bladder cancer, were included in this study. For experiment (1) normal vs. cancer, the LDA classifier on XceptionNet based features provides the best performance with an accuracy of 86.07%, sensitivity of 96.75%, specificity of 69.65%, precision of 83.07%, and F1-score of 89.39%. For experiment (2) non-muscle invasive Cancer (NMIBC) vs. muscle invasive bladder cancer (MIBC), the LDA classifier on XceptionNet based features provided the best performance with an accuracy of 79.72%, sensitivity of 66.62%, specificity of 87.39%, precision of 75.58%, and F1-score of 70.81%. For experiment (3) T0 lesion vs. MIBC, the LDA classifier on XceptionNet based features provides the best performance with an accuracy of 74.96%, sensitivity of 80.51%, specificity of 70.22%, precision of 69.78%, and F1-score of 74.73%. Conclusions: Our proposed model has shown good results in differentiating normal vs cancer and promising performance in differentiating T0 vs MIBC after chemotherapy treatment. We are expanding our dataset to further improve the performance in differentiating T0 vs MIBC. In addition, we will investigate the applicability of GAN for data augmentation to address data limit. We believe the hybrid DL and ML framework may facilitates radiologists' decisions and clinical decision-making in patients with bladder cancer.

Abstract Background Radium (Ra)‐223 is an established treatment option for patients with metastatic castrate‐resistant prostate cancer (mCRPC) who have symptomatic bone metastases without soft tissue disease. Studies have indicated genetic aberrations that regulate DNA damage response (DDR) in prostate cancer can increase susceptibility to treatments such as poly ADP‐ribose polymerase inhibitors and platinum‐based therapies. This study aims to evaluate mCRPC response to Ra‐223 stratified by tumor genomics. Methods This is a retrospective study of mCRPC patients who received Ra‐223 and genetic testing within the Mayo Clinic database (Arizona, Florida, and Minnesota) and Tulane Cancer Center. Patient demographics, genetic aberrations, treatment responses in terms of alkaline phosphatase (ALP) and prostate‐specific antigen (PSA), and survival were assessed. Primary end points were ALP and PSA response. Secondary end points were progression‐free survival (PFS) and overall survival (OS) from time of first radium treatment. Results One hundred and twenty‐seven mCRPC patients treated with Ra‐223 had germline and/or somatic genetic sequencing. The median age at time of diagnosis and Ra‐223 treatment was 61.0 and 68.6 years, respectively. Seventy‐nine (62.2%) had Gleason score ≥ 8 at time of diagnosis. 50.4% received prior docetaxel, and 12.6% received prior cabazitaxel. Notable alterations include TP53 (51.7%), BRCA 1/2 (15.0%), PTEN (13.4%), ATM (11.7%), TMPRSS2‐ERG (8.2%), RB deletion (3.4%), and CDK12 (1.9%). There was no significant difference in ALP or PSA response among the different genetic aberrations. Patients with a TMPRSS2‐ERG mutation exhibited a trend toward lower OS 15.4 months (95% confidence interval [CI] 10.0−NR) versus 26.8 months (95% CI 20.9−35.1). Patients with an RB deletion had a lower PFS 6.0 months (95% CI 1.28−NR) versus 9.0 months (95% CI 7.3−11.1) and a lower OS 13.9 months (95% CI 5.2−NR) versus 26.5 months (95% CI 19.8−33.8). Conclusions Among mCRPC patients treated with Ra‐223 at Mayo Clinic and Tulane Cancer Center, we did not find any clear negative predictors of biochemical response or survival to treatment. TMPRSS2‐ERG and RB mutations were associated with a worse OS. Prospective studies and larger sample sizes are needed to determine the impact of genetic aberrations in response to Ra‐223.

This study was carried out to quantitate the expression levels of microRNA-17, -19a, -34a, -155, and -210 (miRs) expressed in nine clear cell renal cell carcinoma (ccRCC) and one chromophobe renal cell carcinoma cell line with and without sarcomatoid differentiation, and in six primary kidney tumors with matching normal kidney tissues. The data in the five non-sarcomatoid ccRCC cell lines—RC2, CAKI-1, 786-0, RCC4, and RCC4/VHL—and in the four ccRCC with sarcomatoid differentiation—RCJ41T1, RCJ41T2, RCJ41M, and UOK-127—indicated that miR-17 and -19a were expressed at lower levels relative to miR-34a, -155, and -210. Compared with RPTEC normal epithelial cells, miR-34a, miR-155, and miR-210 were expressed at higher levels, independent of the sarcomatoid differentiation status and hypoxia-inducible factors 1α and 2α (HIFs) isoform expression. In the one chromophobe renal cell carcinoma cell line, namely, UOK-276 with sarcomatoid differentiation, and expressing tumor suppressor gene TP53, miR-34a, which is a tumor suppressor gene, was expressed at higher levels than miR-210, -155, -17, and -19a. The pilot results generated in six tumor biopsies with matching normal kidney tissues indicated that while the expression of miR-17 and -19a were similar to the normal tissue expression profile, miR-210, -155, -and 34a were expressed at a higher level. To confirm that differences in the expression levels of the five miRs in the six tumor biopsies were statistically significant, the acquisition of a larger sample size is required. Data previously generated in ccRCC cell lines demonstrating that miR-210, miR-155, and HIFs are druggable targets using a defined dose and schedule of selenium-containing molecules support the concept that simultaneous and concurrent downregulation of miR-210, miR-155, and HIFs, which regulate target genes associated with increased tumor angiogenesis and drug resistance, may offer the potential for the development of a novel mechanism-based strategy for the treatment of patients with advanced ccRCC.

360 Background: Pembrolizumab has been established as an effective treatment option in adjuvant renal cell carcinoma after the publication of KEYNOTE-564. However, the magnitude of benefit based on different risk categories is not well defined. Methods: We included full-text publications of phase II/III randomized controlled trials (RCTs) evaluating immune checkpoint inhibitors or tyrosine kinase inhibitors (TKIs) in adjuvant renal cell carcinoma after a systematic search. The outcomes of interest included disease free survival (DFS), overall survival (OS), all grade or grade ≥3 treatment related, and all cause adverse events. Mixed treatment comparisons were computed through a fixed-effect multivariate meta-regression within the frequent framework. Relative treatment rankings were assessed using P-scores. We used stratified baseline risks of disease progression from observational data and at 5, 10, 15 years from the Leibovich risk stratification system (based on risk scores ranging from 0 to ≥15). Corresponding intervention risks (CIRs) were then approximated using relative effect estimates (from NMA) and baseline risks. The difference between CIRs and baseline risks were calculated to present absolute risk differences in each baseline category. Results: This NMA included six RCTs with a total of 7525 participants and five unique treatment options. Mixed treatment comparisons showed significant DFS and OS benefit with pembrolizumab (rank 1) when compared against sunitinib (DFS HR 0.74 [0.55-0.99]; OS HR 0.51 [0.27-0.94]). However, there were no significant differences with pembrolizumab compared to pazopanib (DFS HR 0.81 [0.60- 1.09]; OS HR 0.54[0.29-1.01]) and axitinib (DFS HR 0.78 [0.54-1.14]; OS HR 0.52 [0.24-1.16]). The safety profiles were comparable. Absolute benefit of TKIs in adjuvant setting was minimal whereas this benefit increased with higher T and N patients in patients treated with pembrolizumab (Table). Similarly, the treatment benefit increased with higher Leibovich’s risk scores at 5, 10 and 15 years of follow up. Conclusions: The current analysis suggests that a risk adapted approach may be useful when considering adjuvant pembrolizumab in RCC patients.[Table: see text]

This Phase Ib study combined programmed death-ligand 1 inhibitor, atezolizumab, with other immunomodulatory agents in locally advanced and metastatic solid tumors. Arms B-D evaluated atezolizumab plus interferon-α, with/without vascular endothelial growth factor inhibitor, bevacizumab, in renal cell carcinoma (RCC) and other solid tumors. Arm B predominantly recruited patients with previously treated RCC or melanoma to receive atezolizumab plus interferon α-2b. Arm C investigated atezolizumab plus polyethylene glycol (PEG)-interferon α-2a in previously treated RCC. Arm D evaluated atezolizumab plus PEG-interferon α-2a and bevacizumab. Primary objectives were safety and tolerability; secondary objectives included clinical activity. Combination therapy was well tolerated, with safety profiles consistent with known risks of individual agents. The most frequent treatment-related toxicities were fatigue, chills, and pyrexia. The objective response rate (ORR) in arm B was 20.0% overall and 17.8% in patients with previously treated checkpoint inhibitor–naive RCC (n = 45). No responses were reported in arm C. The highest ORR in arm D was 46.7% in patients with treatment-naive RCC (n = 15). Data showed preliminary clinical activity and acceptable tolerability of atezolizumab plus interferon α-2b in patients with previously treated checkpoint inhibitor–naive RCC and of atezolizumab plus PEG-interferon α-2a and bevacizumab in patients with treatment-naive RCC.

Abstract Background: Prostate cancers (PC) are intrinsically complex, with few prognostic biomarkers that differentiate aggressive and indolent disease. Cancer associated macrophage like cells (CAMLs) have shown to be significant independent prognostic indicators for poor survival in tumors if engorged to ≥50μm. However, no group has examined CAMLs' in PC. In this prospective pilot study, we analyzed PSA & ≥50μm CAML presence in (n=51) Stage I-III & (n=41) metastatic PC (mPC) prior to the start of new treatments to examine CAMLs' on outcomes. We found that ≥50μm CAML presence significantly predicted worse patient survival in mPC. These preliminary results suggest that CAMLs may serve as a cheap, non-invasive prognostic biomarker that predicts for worse outcomes in both local & mPC prior to new therapies. Methods: We prospectively recruited (n=92) PC patients in this pilot study to examine CAML's prognostic significance in non-metastatic (stages I-III) and mPC. Of (n=92) total patients, 15% (n=14/92) were stage I, 32% (n=29/92) stage II, 9% (n=8/92) stage III, and 45% (n=41/92) Stage IV. Prior to the induction of new therapy or first line therapy for newly diagnosed patients, 7.5mL peripheral baseline (BL) blood was collected. Blood was filtered via CellSieveTM microfiltration, which excludes cells by size ≥7μm. In parallel, changes in PSA was monitored to compare prognostic significance against ≥50μm CAMLs. Wilcoxon univariate analysis was used to analyze ≥50μm CAMLs and PSA in progression free survival (PFS) and overall survival (OS). Results: CAML's were identified in 78.9% (n=71/90) of available BL samples, with two samples failing from clotting. Stage I-III patients had an average of ~3 CAMLs/7.5mL blood, whereas mPCs contained ~6 CAMLs/7.5mL. Average CAML size increased linearly with advancing stage of disease with stage I patients averaging 23μm, stage II 33μm, stage III 65μm, and stage IV 78μm. Of available BL samples, 41% (n=38/90) patients with ≥50μm CAMLs significantly predicted for worse PFS (HR=10.0, p<0.001, 95%CI=0.07-0.31) and OS (HR=34.2, p<0.001, 95%CI=0.03-0.21). Further, ≥50μm CAMLs significantly predicted for poorer survival in mPC with PFS (HR=9.4, p=0.028, 95%CI=0.18-0.84) and OS (HR=5.1, p=0.003, 95%CI=0.07-0.53), but only non-significantly trended in non-metastatic disease. Initial analysis of PSA found that high PSA at BL was a significant predictor for worse outcomes in patients, while changes in PSA was not a predictor. Conclusions: CAMLs have shown potential as a predictive non-invasive blood based biomarker across multiple solid malignancies, and here we compared this biomarker in PC. While preliminary, these data suggests that engorged CAML presence prior to the start of new treatment may be a statistically significant predictor of survival in both the local and metastatic setting, and a larger validation cohort should be established. Citation Format: Daniel J. Gironda, Raymond Bergan, R K. Alpaugh, Daniel C. Danila, Tuan L. Chuang, Brenda Y. Hurtado, Thai Ho, Cha-Mei Tang, Daniel L. Adams. Cancer associated macrophage like cells predict aggressive disease in local and metastatic prostate cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 357.

3039 Background: Patients with multiple organ metastases have poorer prognoses than those with a single organ metastasis, are frequently associated with drug resistance, and have higher tumor burden. Engorged (≥50um) Cancer Associated Macrophage-Like Cells (CAMLs) are a circulating stromal cell subtype detected in the blood of patients with solid tumors at high risk for recurrence or progression. While numerous studies have shown that ≥50um CAMLs predict poor clinical outcomes, meta-analysis of these studies have also suggested that hyper engorged CAMLs ≥100um (heCAMLs) may be associated with multifocal metastatic disease and even worse outcomes. In this prospective study, we evaluated the presence of heCAMLs in patients with metastatic disease and demonstrated a strong relationship with multi organ spread, which also correlated with shorter Progression Free Survival (PFS) and Overall Survival (OS). Methods: We prospectively recruited 151 patients with metastatic (m) mbreast (n = 58), mlung (n = 34), mprostate (n = 39), and mrenal (n = 20) cancers. Peripheral blood was collected prior to the induction of new treatment for metastatic cancer. Cells were isolated following standard CellSieve techniques, then imaged and measured in ZenBlue. Multi organ metastasis was defined as spread to ≥2 distant organ sites, or any spread to the brain. Single factor ANOVA was conducted to compare heCAML presence in multi organ metastatic patients versus patients with single organ site metastasis. Univariate and multivariate analysis was run to evaluate for PFS and OS against heCAMLs, and all known clinical parameters. Results: 150 viable samples (excluding 1 failed sample) were obtained. Multi organ metastases were present in 55% (n = 83/150) of patients. heCAMLs were found in 59% (n = 49/83) of the multi organ metastatic population, but only in 16% (n = 11/67) of the single site metastatic cohort (p < 0.001). heCAML presence appeared to differentiate multi organ vs single organ metastases in mbreast (85% vs. 52%, p = 0.006), mlung (71% vs. 26%, p = 0.025), mprostate (75% vs. 37%, p = 0.029), and mRCC (88% vs. 36%, p = 0.025). Further, in all n = 150 patients, heCAML presence predicted a significantly shorter median PFS of 4.5 versus 7.2 months, 24 month PFS (HR = 1.67, 95%CI = 1.13-2.45, p = 0.013), and significantly shorter median OS of 13.1 versus 20.4 months, 24 month OS (HR = 2.05, 95%CI = 1.24-3.39, p = 0.008). Conclusions: We examined a non-invasive prognostic blood based assay to determine its relationship to multi organ metastatic spread as well as its prognostic value in several solid cancers. These results showed patients with heCAMLs had higher rates of multi organ metastases, and appear to predict for shorter PFS and OS. Studies of larger cohorts are needed for prospective validation of these initial findings.