Urachal and Nonurachal Adenocarcinomas of the Urinary Bladder: A Comprehensive Genomic Profiling Study.

PURPOSE: Although both urachal (U) and nonurachal (NU) bladder adenocarcinomas (adenoCas) share several histologic similarities, they differ in location and sometimes in therapeutic options. We analyzed the differences in genomic alterations (GAs) between these tumor entities, with the aim of identifying potential therapeutic targets for clinical trials. MATERIALS AND METHODS: Overall, 133 U and 328 NU adenoCas were analyzed. Hybrid capture-based comprehensive genomic profiling (CGP) was performed to evaluate all classes of GA. Germline status of GA was predicted using a validated somatic-germline computational method. CGP was performed using the FoundationOne and FoundationOne CDx assays (Foundation Medicine, Inc). RESULTS: The most frequent GA in both U and NU cohorts included TP53 (86.5% v 81.1%) and KRAS (34.6% v 27.7%). GAs characteristic of colorectal adenoCa, such as SMAD4 (P = .069) and GNAS (P = .071), were more common in U versus NU. Conversely, TERT (P < .01) and RB1 (P = .071) were more prevalent in NU adenoCa. Notably, both U and NU adenoCas exhibited possibly targetable GA in PIK3CA (7.5% v 7.9%) and ERBB2 (6.8% v 7.6%). Biomarkers associated with potential benefit from anti-PD-1/L1 were infrequent. Median tumor mutational burden was 2.6 and 3.5 mutations per megabase for U and NU, respectively, and PD-L1 expression >1% was rare. Genomic ancestry and genomic signature distribution were similar in both tumor types. GAs were most commonly of somatic nature. Limitations include lack of clinical data, tumor heterogeneity, and retrospective nature. CONCLUSION: U and NU adenoCAs revealed differences in GA, with PIK3CA and ERBB2 being identified as putative therapeutic targets. Biomarkers of response to anti-PD-(L)1 were uncommon. Results highlight the potential of CGP to personalize treatment options of bladder adenoCa and inform clinical trial designs.

Enhancing cancer prediction in challenging screen-detected incident lung nodules using time-series deep learning.

Lung cancer screening (LCS) using annual computed tomography (CT) scanning significantly reduces mortality by detecting cancerous lung nodules at an earlier stage. Deep learning algorithms can improve nodule malignancy risk stratification. However, they have typically been used to analyse single time point CT data when detecting malignant nodules on either baseline or incident CT LCS rounds. Deep learning algorithms have the greatest value in two aspects. These approaches have great potential in assessing nodule change across time-series CT scans where subtle changes may be challenging to identify using the human eye alone. Moreover, they could be targeted to detect nodules developing on incident screening rounds, where cancers are generally smaller and more challenging to detect confidently. Here, we show the performance of our Deep learning-based Computer-Aided Diagnosis model integrating Nodule and Lung imaging data with clinical Metadata Longitudinally (DeepCAD-NLM-L) for malignancy prediction. DeepCAD-NLM-L showed improved performance (AUC = 88%) against models utilizing single time-point data alone. DeepCAD-NLM-L also demonstrated comparable and complementary performance to radiologists when interpreting the most challenging nodules typically found in LCS programs. It also demonstrated similar performance to radiologists when assessed on out-of-distribution imaging dataset. The results emphasize the advantages of using time-series and multimodal analyses when interpreting malignancy risk in LCS.

Oncolytic adenoviral therapy plus pembrolizumab in BCG-unresponsive non-muscle-invasive bladder cancer: the phase 2 CORE-001 trial.

Cretostimogene grenadenorepvec is a serotype-5 oncolytic adenovirus designed to selectively replicate in cancer cells with retinoblastoma pathway alterations, previously tested as monotherapy in bacillus Calmette-Guérin (BCG)-experienced non-muscle-invasive bladder cancer. In this phase 2 study, we assessed the potential synergistic efficacy between intravesical cretostimogene and systemic pembrolizumab in patients with BCG-unresponsive non-muscle-invasive bladder cancer with carcinoma in situ (CIS). Thirty-five patients were treated with intravesical cretostimogene with systemic pembrolizumab. Induction cretostimogene was administered weekly for 6 weeks followed by three weekly maintenance infusions at months 3, 6, 9, 12 and 18 in patients maintaining complete response (CR). Patients with persistent CIS/high-grade Ta at the 3-month assessment were eligible for re-induction. Pembrolizumab was administered for up to 24 months. The primary endpoint was CR at 12 months as assessed by cystoscopy, urine cytology, cross-sectional imaging and mandatory bladder mapping biopsies. Secondary endpoints included CR at any time, duration of response, progression-free survival and safety. The CR rate in the intention-to-treat population at 12 months was 57.1% (20 out of 35, 95% confidence interval (CI) 40.7-73.5%), meeting the primary endpoint. A total of 29 out of 35 patients (82.9%, 95% CI 70.4-95.3%) derived a CR at 3 months. With a median follow-up of 26.5 months, the median duration of response has not been reached (95% CI 15.7 to not reached). The CR rate at 24 months was 51.4% (18 out of 35) (95% CI 34.9-68.0%). No patient progressed to muscle-invasive bladder cancer in this trial. Adverse events attributed to cretostimogene were low grade, self-limiting and predominantly limited to bladder-related symptoms. A total of 5 out of 35 patients (14.3%) developed grade 3 treatment-related adverse effects. There was no evidence of overlapping or synergistic toxicities. Combination intravesical cretostimogene and systemic pembrolizumab demonstrated enduring efficacy. With a toxicity profile similar to its monotherapy components, this combination may shift the benefit-to-risk ratio for patients with BCG-unresponsive CIS. ClinicalTrials.gov Identifier: NCT04387461 .

Delineating excess comorbidities in idiopathic pulmonary fibrosis: an observational study.

BACKGROUND: Our study examined whether prevalent and incident comorbidities are increased in idiopathic pulmonary fibrosis (IPF) patients when compared to matched chronic obstructive pulmonary disease (COPD) patients and control subjects without IPF or COPD. METHODS: IPF and age, gender and smoking matched COPD patients, diagnosed between 01/01/1997 and 01/01/2019 were identified from the Clinical Practice Research Datalink GOLD database multiple registrations cohort at the first date an ICD-10 or read code mentioned IPF/COPD. A control cohort comprised age, gender and pack-year smoking matched subjects without IPF or COPD. Prevalent (prior to IPF/COPD diagnosis) and incident (after IPF/COPD diagnosis) comorbidities were examined. Group differences were estimated using a t-test. Mortality relationships were examined using multivariable Cox proportional hazards adjusted for patient age, gender and smoking status. RESULTS: Across 3055 IPF patients, 38% had 3 or more prevalent comorbidities versus 32% of COPD patients and 21% of matched control subjects. Survival time reduced as the number of comorbidities in an individual increased (p < 0.0001). In IPF, prevalent heart failure (Hazard ratio [HR] = 1.62, 95% Confidence Interval [CI]: 1.43-1.84, p < 0.001), chronic kidney disease (HR = 1.27, 95%CI: 1.10-1.47, p = 0.001), cerebrovascular disease (HR = 1.18, 95%CI: 1.02-1.35, p = 0.02), abdominal and peripheral vascular disease (HR = 1.29, 95%CI: 1.09-1.50, p = 0.003) independently associated with reduced survival. Key comorbidities showed increased incidence in IPF (versus COPD) 7-10 years prior to IPF diagnosis. INTERPRETATION: The mortality impact of excessive prevalent comorbidities in IPF versus COPD and smoking matched controls suggests that multiorgan mechanisms of injury need elucidation in patients that develop IPF.

Frequency and Nature of Genomic Alterations in ERBB2-Altered Urothelial Bladder Cancer.

BACKGROUND: Human epidermal growth factor-2 (HER2) overexpression is an oncogenic driver in many solid tumors, including urothelial bladder cancer (UBC). In addition, activating mutations in the ERBB2 gene have been shown to play an oncogenic role similar to ERBB2 amplification. OBJECTIVE: To describe and compare the frequency and nature of genomic alterations (GA) of ERBB2-altered (mutations, amplification) and ERBB2 wild-type UBC. PATIENTS AND METHODS: Using a hybrid capture-based comprehensive profiling assay, 9518 UBC cases were grouped by ERBB2 alteration and evaluated for all classes of genomic alterations (GA), tumor mutational burden (TMB), microsatellite instability (MSI), genome-wide loss of heterozygosity (gLOH), and genomic mutational signature. PD-L1 expression was measured by immunohistochemistry (Dako 22C3). Categorical statistical comparisons were performed using Fisher's exact tests. RESULTS: A total of 602 (6.3%) UBC cases featured ERBB2 extracellular domain short variant (SV) GA (ECDmut+), 253 (2.7%) cases featured ERBB2 kinase domain SV GA (KDmut+), 866 (9.1%) cases had ERBB2 amplification (amp+), and 7797 (81.9%) cases were ERBB2 wild-type (wt). European genetic ancestry of ECDmut+ was higher than ERBB2wt. Numerous significant associations were observed when comparing GA by group. Notably among these, CDKN2A/MTAP loss were more frequent in ERBB2wt versus ECDmut+ and amp+. ERBB3 GA were more frequent in ECDmut+ and KDmut+ than ERBB2wt. TERT GA were more frequent in ECDmut+, KDmut+, and amp+ versus ERBB2wt. TOP2A amplification was significantly more common in ECDmut+ and amp+ versus ERBB2wt, and TP53 SV GA were significantly higher in ERBB2 amp+ versus ERBB2wt. Mean TMB levels were significantly higher in ECDmut+, KDmut+, and amp+ than in ERBB2wt. Apolipoprotein B mRNA-editing enzyme, catalytic polypeptides (APOBEC) signature was more frequent in ECDmut+, KDmut+, and amp+ versus ERBB2wt. No significant differences were observed in PD-L1 status between groups, while gLOH-high status was more common in amp+ versus ERBB2wt. MSI-high status was more frequent in KDmut+ versus ERBB2wt, and in ERBB2wt than in amp+. CONCLUSIONS: We noted important differences in co-occurring GA in ERBB2-altered (ECDmut+, KDmut+, amp+) versus ERBB2wt UBC, as well as higher mean TMB and higher APOBEC mutational signature in the ERBB2-altered groups. Our results can help refine future clinical trial designs and elucidate possible response and resistance mechanisms for ERBB2-altered UBC.

Expectation maximisation pseudo labels.

In this paper, we study pseudo-labelling. Pseudo-labelling employs raw inferences on unlabelled data as pseudo-labels for self-training. We elucidate the empirical successes of pseudo-labelling by establishing a link between this technique and the Expectation Maximisation algorithm. Through this, we realise that the original pseudo-labelling serves as an empirical estimation of its more comprehensive underlying formulation. Following this insight, we present a full generalisation of pseudo-labels under Bayes' theorem, termed Bayesian Pseudo Labels. Subsequently, we introduce a variational approach to generate these Bayesian Pseudo Labels, involving the learning of a threshold to automatically select high-quality pseudo labels. In the remainder of the paper, we showcase the applications of pseudo-labelling and its generalised form, Bayesian Pseudo-Labelling, in the semi-supervised segmentation of medical images. Specifically, we focus on: (1) 3D binary segmentation of lung vessels from CT volumes; (2) 2D multi-class segmentation of brain tumours from MRI volumes; (3) 3D binary segmentation of whole brain tumours from MRI volumes; and (4) 3D binary segmentation of prostate from MRI volumes. We further demonstrate that pseudo-labels can enhance the robustness of the learned representations. The code is released in the following GitHub repository: https://github.com/moucheng2017/EMSSL.

The effect of dietary interventions or patterns on the cardiometabolic health of individuals treated with androgen deprivation therapy for prostate cancer: A systematic review.

Prostate cancer survivors treated with androgen deprivation therapy may be at increased risk of cardiovascular disease. Dietary recommendations for the prevention and/or management of cardiovascular disease for these individuals are lacking. This review synthesizes the evidence on the effect of dietary interventions on cardiometabolic biomarkers and cardiovascular disease risk in prostate cancer survivors receiving androgen deprivation therapy. A systematic review was conducted across PubMed, CINAHL, Embase, and Cochrane CENTRAL. Intervention or observational cohort studies evaluating diets, nutrients, or nutraceuticals with or without concurrent exercise interventions on cardiovascular disease, cardiovascular events, or cardiovascular disease biomarkers in those treated with androgen deprivation therapy were included. Confidence in the body of evidence was appraised using Grading of Recommendations, Assessment, Development and Evaluations. Twelve studies reported across fifteen papers were included. Interventions were heterogenous, with most studies including an exercise co-intervention (n = 8). Few significant findings for the effects of diet on cardiometabolic markers were likely due to weak methodology and sample sizes. Strongest evidence was for the effect of a healthy Western dietary pattern with exercise on improved blood pressure (Confidence: moderate). The healthy Western dietary pattern with exercise may improve high-density lipoprotein cholesterol (Confidence: Low) and flow-mediated dilation. Soy may improve total cholesterol (Confidence: Very low). A low-carbohydrate diet with physical activity may improve high-density lipoprotein cholesterol, incidence of metabolic syndrome, and Framingham cardiovascular disease risk score. Evidence of the effect of dietary interventions on cardiometabolic biomarkers and cardiovascular disease risk of prostate cancer survivors receiving androgen deprivation therapy is insufficient to inform practice. Well-designed dietary interventions aimed at improving cardiometabolic outcomes of this population are warranted to inform future dietary recommendations.

Characterizing the Genomic Landscape of the Micropapillary Subtype of Urothelial Carcinoma of the Bladder Harboring Activating Extracellular Mutations of ERBB2.

The micropapillary subtype of urothelial carcinoma (MPUC) of the bladder is a very aggressive histological variant of urothelial bladder cancer (UBC). A high frequency of MPUC contains activating mutations in the extracellular domain (ECD) of ERBB2. We sought to further characterize ERBB2 ECD-mutated MPUC to identify additional genomic alterations that have been associated with tumor progression and therapeutic response. In total, 5,485 cases of archived formalin-fixed, paraffin-embedded UBC underwent comprehensive genomic profiling to identify ERBB2 ECD-mutated MPUC and evaluate the frequencies of genomic co-alterations. We identified 219 cases of UBC with ERBB2 ECD mutations (74% S310F and 26% S310Y), of which 63 (28.8%) were MPUC. Genomic analysis revealed that TERT, TP53, and ARID1A were the most common co-altered genes in ERBB2-mutant MPUC (82.5%, 58.7%, and 39.7%, respectively) and did not differ from ERBB2-mutant non-MPUC (86.5%, 51.9%, and 35.3%). The main differences between ERBB2 ECD-mutated MPUC compared with non-MPUC were KMT2D, RB1, and MTAP alterations. KMT2D and RB1 are tumor-suppressor genes. KMT2D frequency was significantly decreased in ERBB2 ECD-mutated MPUC (6.3%) in contrast to non-MPUC (27.6%; P < .001). RB1 mutations were more frequent in ERBB2 ECD-mutated MPUC (33.3%) than in non-MPUC (17.3%; P = .012). Finally, MTAP loss, an emerging biomarker for new synthetic lethality-based anticancer drugs, was less frequent in ERBB2 ECD-mutated MPUC (11.1%) than in non-MPUC (26.9%; P = .018). Characterizing the genomic landscape of MPUC may not only improve our fundamental knowledge about this aggressive morphological variant of UBC but also has the potential to identify possible prognostic and predictive biomarkers that may drive tumor progression and dictate treatment response to therapeutic approaches.

A formula for predicting emphysema extent in combined idiopathic pulmonary fibrosis and emphysema.

BACKGROUND: No single pulmonary function test captures the functional effect of emphysema in idiopathic pulmonary fibrosis (IPF). Without experienced radiologists, other methods are needed to determine emphysema extent. Here, we report the development and validation of a formula to predict emphysema extent in patients with IPF and emphysema. METHODS: The development cohort included 76 patients with combined IPF and emphysema at the Royal Brompton Hospital, London, United Kingdom. The formula was derived using stepwise regression to generate the weighted combination of pulmonary function data that fitted best with emphysema extent on high-resolution computed tomography. Test cohorts included patients from two clinical trials (n = 455 [n = 174 with emphysema]; NCT00047645, NCT00075998) and a real-world cohort from the Royal Brompton Hospital (n = 191 [n = 110 with emphysema]). The formula is only applicable for patients with IPF and concomitant emphysema and accordingly was not used to detect the presence or absence of emphysema. RESULTS: The formula was: predicted emphysema extent = 12.67 + (0.92 x percent predicted forced vital capacity) - (0.65 x percent predicted forced expiratory volume in 1 second) - (0.52 x percent predicted carbon monoxide diffusing capacity). A significant relationship between the formula and observed emphysema extent was found in both cohorts (R2 = 0.25, P < 0.0001; R2 = 0.47, P < 0.0001, respectively). In both, the formula better predicted observed emphysema extent versus individual pulmonary function tests. A 15% emphysema extent threshold, calculated using the formula, identified a significant difference in absolute changes from baseline in forced vital capacity at Week 48 in patients with baseline-predicted emphysema extent < 15% versus ≥ 15% (P = 0.0105). CONCLUSION: The formula, designed for use in patients with IPF and emphysema, demonstrated enhanced ability to predict emphysema extent versus individual pulmonary function tests. TRIAL REGISTRATION: NCT00047645; NCT00075998.

Imaging of pulmonary fibrosis in children: A review, with proposed diagnostic criteria.

Computed tomography (CT) imaging findings of pulmonary fibrosis are well established for adults and have been shown to correlate with prognosis and outcome. Recognition of fibrotic CT findings in children is more limited. With approved treatments for adult pulmonary fibrosis, it has become critical to define CT criteria for fibrosis in children, to identify patients in need of treatment and those eligible for clinical trials. Understanding how pediatric fibrosis compares with idiopathic pulmonary fibrosis and other causes of fibrosis in adults is increasingly important as these patients transition to adult care teams. Here, we review what is known regarding the features of pulmonary fibrosis in children compared with adults. Pulmonary fibrosis in children may be associated with genetic surfactant dysfunction disorders, autoimmune systemic disorders, and complications after radiation, chemotherapy, transplantation, and other exposures. Rather than a basal-predominant usual interstitial pneumonia pattern with honeycombing, pediatric fibrosis is primarily characterized by reticulation, traction bronchiectasis, architectural distortion, or cystic lucencies/abnormalities. Ground-glass opacities are more frequent in children with fibrotic interstitial lung disease than adults, and disease distribution appears more diffuse, without clearly defined axial or craniocaudal predominance. Following discussion and consensus amongst a panel of expert radiologists, pathologists and physicians, distinctive disease features were integrated to develop criteria for the first global Phase III trial in children with pulmonary fibrosis.

Serum C-reactive protein is associated with earlier mortality across different interstitial lung diseases.

BACKGROUND AND OBJECTIVE: The acute-phase protein C-reactive protein (CRP) is known to be associated with poor outcomes in cancer and cardiovascular disease, but there is limited evidence of its prognostic implications in interstitial lung diseases (ILDs). We therefore set out to test whether baseline serum CRP levels are associated with mortality in four different ILDs. METHODS: In this retrospective study, clinically measured CRP levels, as well as baseline demographics and lung function measures, were collected for ILD patients first presenting to the Royal Brompton Hospital between January 2010 and December 2019. Cox regression analysis was used to determine the relationship with 5-year mortality. RESULTS: Patients included in the study were: idiopathic pulmonary fibrosis (IPF) n = 422, fibrotic hypersensitivity pneumonitis (fHP) n = 233, rheumatoid arthritis associated ILD (RA-ILD) n = 111 and Systemic Sclerosis associated ILD (SSc-ILD) n = 86. Patients with a recent history of infection were excluded. Higher CRP levels were associated with shorter 5-year survival in all four disease groups on both univariable analyses, and after adjusting for age, gender, smoking history, immunosuppressive therapy and baseline disease severity (IPF: HR (95% CI): 1.3 (1.1-1.5), p = 0.003, fHP: 1.5 (1.2-1.9), p = 0.001, RA-ILD: 1.4 (1.1-1.84), p = 0.01 and SSc-ILD: 2.7 (1.6-4.5), p < 0.001). CONCLUSION: Higher CRP levels are independently associated with reduced 5-year survival in IPF, fHP, RA-ILD and SSc-ILD.

Mortality surrogates in combined pulmonary fibrosis and emphysema.

BACKGROUND: Idiopathic pulmonary fibrosis (IPF) with coexistent emphysema, termed combined pulmonary fibrosis and emphysema (CPFE) may associate with reduced forced vital capacity (FVC) declines compared to non-CPFE IPF patients. We examined associations between mortality and functional measures of disease progression in two IPF cohorts. METHODS: Visual emphysema presence (>0% emphysema) scored on computed tomography identified CPFE patients (CPFE/non-CPFE: derivation cohort n=317/n=183, replication cohort n=358/n=152), who were subgrouped using 10% or 15% visual emphysema thresholds, and an unsupervised machine-learning model considering emphysema and interstitial lung disease extents. Baseline characteristics, 1-year relative FVC and diffusing capacity of the lung for carbon monoxide (D LCO) decline (linear mixed-effects models), and their associations with mortality (multivariable Cox regression models) were compared across non-CPFE and CPFE subgroups. RESULTS: In both IPF cohorts, CPFE patients with ≥10% emphysema had a greater smoking history and lower baseline D LCO compared to CPFE patients with <10% emphysema. Using multivariable Cox regression analyses in patients with ≥10% emphysema, 1-year D LCO decline showed stronger mortality associations than 1-year FVC decline. Results were maintained in patients suitable for therapeutic IPF trials and in subjects subgrouped by ≥15% emphysema and using unsupervised machine learning. Importantly, the unsupervised machine-learning approach identified CPFE patients in whom FVC decline did not associate strongly with mortality. In non-CPFE IPF patients, 1-year FVC declines ≥5% and ≥10% showed strong mortality associations. CONCLUSION: When assessing disease progression in IPF, D LCO decline should be considered in patients with ≥10% emphysema and a ≥5% 1-year relative FVC decline threshold considered in non-CPFE IPF patients.

First analysis of the safety and efficacy of UGN-101 in the treatment of ureteral tumors.

OBJECTIVE: UGN-101 has been approved for the chemoablation of low-grade upper tract urothelial cancer (UTUC) involving the renal pelvis and calyces. Herein is the first reported cohort of patients with ureteral tumors treated with UGN-101. PATIENTS AND METHODS: We performed a retrospective review of patients treated with UGN-101 for UTUC at 15 high-volume academic and community centers focusing on outcomes of patients treated for ureteral disease. Patients received UGN-101 with either adjuvant or chemo-ablative intent. Response rates are reported for patients receiving chemo-ablative intent. Adverse outcomes were characterized with a focus on the rate of ureteral stenosis. RESULTS: In a cohort of 132 patients and 136 renal units, 47 cases had tumor involvement of the ureter, with 12 cases of ureteral tumor only (8.8%) and 35 cases of ureteral plus renal pelvic tumors (25.7%). Of the 23 patients with ureteral involvement who received UGN-101 induction with chemo-ablative intent, the complete response was 47.8%, which did not differ significantly from outcomes in patients without ureteral involvement. Fourteen patients (37.8%) with ureteral tumors had significant ureteral stenosis at first post-treatment evaluation, however, when excluding those with pre-existing hydronephrosis or ureteral stenosis, only 5.4% of patients developed new clinically significant stenosis. CONCLUSIONS: UGN-101 appears to be safe and may have similar efficacy in treating low-grade urothelial carcinoma of the ureter as compared to renal pelvic tumors.

CenTime: Event-conditional modelling of censoring in survival analysis.

Survival analysis is a valuable tool for estimating the time until specific events, such as death or cancer recurrence, based on baseline observations. This is particularly useful in healthcare to prognostically predict clinically important events based on patient data. However, existing approaches often have limitations; some focus only on ranking patients by survivability, neglecting to estimate the actual event time, while others treat the problem as a classification task, ignoring the inherent time-ordered structure of the events. Additionally, the effective utilisation of censored samples-data points where the event time is unknown- is essential for enhancing the model's predictive accuracy. In this paper, we introduce CenTime, a novel approach to survival analysis that directly estimates the time to event. Our method features an innovative event-conditional censoring mechanism that performs robustly even when uncensored data is scarce. We demonstrate that our approach forms a consistent estimator for the event model parameters, even in the absence of uncensored data. Furthermore, CenTime is easily integrated with deep learning models with no restrictions on batch size or the number of uncensored samples. We compare our approach to standard survival analysis methods, including the Cox proportional-hazard model and DeepHit. Our results indicate that CenTime offers state-of-the-art performance in predicting time-to-death while maintaining comparable ranking performance. Our implementation is publicly available at https://github.com/ahmedhshahin/CenTime.

Genomic Profiles and Clinical Outcomes of Penile Squamous Cell Carcinoma With Elevated Tumor Mutational Burden.

Importance: Tumor mutational burden (TMB) is a putative biomarker of efficacy for immune checkpoint inhibitor (ICI) therapies of solid tumors, but not specifically for penile squamous cell carcinoma (PSCC). Objective: To characterize biomarker features and ICI therapy outcomes associated with high TMB in PSCC in the routine clinical practice setting. Design, Setting, and Participants: In this cohort study, 397 PSCC cases were analyzed to identify genomic alterations in more than 300 cancer-associated genes and genomic signatures, including TMB, using a hybrid capture-based comprehensive genomic profiling assay. Tumor mutational burden was categorized as low (<10 mutations per megabase [mut/Mb]), high (10-19 mut/Mb), or very high (≥20 mut/Mb). Germline status of genetic alterations was predicted using a validated somatic-germline computational method. Clinical outcomes of patients with metastatic PSCC receiving first-line ICI were abstracted using the deidentified nationwide Clinico-Genomic Database (CGDB) from January 1, 2011, through December 31, 2022. Exposure: Comprehensive genomic profiling was performed using FoundationOne and FoundationOne CDx assays from Foundation Medicine Inc. Main outcomes and measures: The spectrum of genetic alterations by TMB level in PSCC, the percentage of germline genetic alterations, and the outcome (overall survival with routine clinical treatment) by TMB of chemotherapy-naive patients with PSCC who received ICI treatment up front were assessed in this descriptive study. Results: Among 397 patients (median [IQR] age, 65 [54-73] years; 266 [67.0%] of European, 83 [20.9%] of admixed American, and 34 [8.5%] of African or other genomic ancestry), the median (IQR) age (eg, 65 [53-73] years for low TMB vs 68 [61-78] years for TMB ≥10 mut/Mb) and genomic ancestry distribution (eg, European 228 of 339 [67.3%] for low TMB vs 38 of 58 [65.5%] for TMB ≥10 mut/Mb) were similar between TMB subgroups. There were 339 PSCC cases (85.4%) with low TMB, 40 cases (10.1%) with high TMB, and 18 cases (4.5%) with very high TMB. Comparisons of TMB of 10 mut/Mb or higher vs low TMB showed an enrichment of genetic alterations in PIK3CA (48.3% vs 18.3%; P < .001) and KMT2D (29.3% vs 7.7%; P < .001) and less frequent genetic alterations in CDKN2A (25.9% vs 45.7%; P = .05). Most genetic alterations did not co-occur. Human papillomavirus identification was more frequent as TMB increased: 28.3% for low TMB, 50.0% for high, and 72.2% for very high. In total, 95 of 1377 genetic alterations (6.9%) were germline. Of 10 patients identified from the CGDB receiving frontline ICIs, median (IQR) follow-up was 9.9 months. Four patients had overall survival with clinical treatment of more than 12 months, including 2 of 3 patients with TMB of 10 mut/Mb or higher. Conclusions and Relevance: In this cohort study of advanced metastatic PSCC based on TMB levels, significant differences were observed for biomarkers in nearly 15% of patients with a TMB of 10 mut/Mb or higher. Germline testing and ICI-based therapy should be integrated into the management of selected PSCC cases.

Learning from multiple annotators for medical image segmentation.

Supervised machine learning methods have been widely developed for segmentation tasks in recent years. However, the quality of labels has high impact on the predictive performance of these algorithms. This issue is particularly acute in the medical image domain, where both the cost of annotation and the inter-observer variability are high. Different human experts contribute estimates of the "actual" segmentation labels in a typical label acquisition process, influenced by their personal biases and competency levels. The performance of automatic segmentation algorithms is limited when these noisy labels are used as the expert consensus label. In this work, we use two coupled CNNs to jointly learn, from purely noisy observations alone, the reliability of individual annotators and the expert consensus label distributions. The separation of the two is achieved by maximally describing the annotator's "unreliable behavior" (we call it "maximally unreliable") while achieving high fidelity with the noisy training data. We first create a toy segmentation dataset using MNIST and investigate the properties of the proposed algorithm. We then use three public medical imaging segmentation datasets to demonstrate our method's efficacy, including both simulated (where necessary) and real-world annotations: 1) ISBI2015 (multiple-sclerosis lesions); 2) BraTS (brain tumors); 3) LIDC-IDRI (lung abnormalities). Finally, we create a real-world multiple sclerosis lesion dataset (QSMSC at UCL: Queen Square Multiple Sclerosis Center at UCL, UK) with manual segmentations from 4 different annotators (3 radiologists with different level skills and 1 expert to generate the expert consensus label). In all datasets, our method consistently outperforms competing methods and relevant baselines, especially when the number of annotations is small and the amount of disagreement is large. The studies also reveal that the system is capable of capturing the complicated spatial characteristics of annotators' mistakes.

A hybrid CNN-RNN approach for survival analysis in a Lung Cancer Screening study.

In this study, we present a hybrid CNN-RNN approach to investigate long-term survival of subjects in a lung cancer screening study. Subjects who died of cardiovascular and respiratory causes were identified whereby the CNN model was used to capture imaging features in the CT scans and the RNN model was used to investigate time series and thus global information. To account for heterogeneity in patients' follow-up times, two different variants of LSTM models were evaluated, each incorporating different strategies to address irregularities in follow-up time. The models were trained on subjects who underwent cardiovascular and respiratory deaths and a control cohort matched to participant age, gender, and smoking history. The combined model can achieve an AUC of 0.76 which outperforms humans at cardiovascular mortality prediction. The corresponding F1 and Matthews Correlation Coefficient are 0.63 and 0.42 respectively. The generalisability of the model is further validated on an 'external' cohort. The same models were applied to survival analysis with the Cox Proportional Hazard model. It was demonstrated that incorporating the follow-up history can lead to improvement in survival prediction. The Cox neural network can achieve an IPCW C-index of 0.75 on the internal dataset and 0.69 on an external dataset. Delineating subjects at increased risk of cardiorespiratory mortality can alert clinicians to request further more detailed functional or imaging studies to improve the assessment of cardiorespiratory disease burden. Such strategies may uncover unsuspected and under-recognised pathologies thereby potentially reducing patient morbidity.