Distribution and Positive Predictive Value of Follicle Stimulating Hormone Among Nonazoospermic Men.

PURPOSE: Our goal was to characterize the distribution of follicle stimulating hormone (FSH) in fertile and subfertile nonazoospermic men, and to determine the ability of various FSH thresholds to predict fertility status. MATERIALS AND METHODS: We performed a retrospective cohort study of 1389 nonazoospermic men who presented for fertility evaluation. Men with at least 2 semen analyses and 1 FSH level were included. Men were dichotomized into fertile and subfertile groups based on total motile sperm count. FSH was evaluated within a multivariable model, and positive predictive values (PPVs) for subfertility were used to assess the clinical utility of various FSH thresholds. RESULTS: One thousand fifteen (80%) men were classified as fertile and 274 (20%) as subfertile. Age, presence of varicocele, and testosterone levels were not statistically different between the groups. Median FSH was 4.0 vs 6.0 (P < .001) among fertile vs subfertile men. Multiple FSH thresholds ranging from 2.9 to 9.3 performed similarly in predicting fertility status (PPV 0.49-0.59). Only FSH thresholds above the 95th percentile (12.1) had PPVs greater than 0.7. The highest PPV (0.84) was seen at an FSH of 20.8 (99th percentile). CONCLUSIONS: While there were significant differences in FSH levels among fertile and subfertile nonazoospermic men, multiple FSH cutoffs between 2.2 and 9.3 performed poorly for prediction of fertility status as determined by total motile sperm count. It was not until the 95th percentile FSH value that a clinically useful level of predictability for subfertility was reached, indicating that FSH should not be used as a standalone test of fertility status. Nonetheless, FSH testing remains clinically useful and may be most informative in the setting of extreme values or discordant FSH and semen analysis results.

Combination of pembrolizumab and pelareorep promotes anti-tumour immunity in advanced pancreatic adenocarcinoma (PDAC).

BACKGROUND: We previously reported activity of pelareorep, pembrolizumab and chemotherapy. Patients developed new T-cell clones and increased peripheral T-cell clonality, leading to an inflamed tumour. To evaluate a chemotherapy-free regimen, this study assesses if pelareorep and pembrolizumab has efficacy by inducing anti-tumour immunological changes (NCT03723915). METHODS: PDAC patients who progressed after first-line therapy, received iv pelareorep induction with pembrolizumab every 21-days. Primary objective is overall response rate. Secondary objectives included evaluation of immunological changes within tumour and blood. RESULTS: Clinical benefit rate (CBR) was 42% amongst 12 patients. One patient achieved partial response (PR) and four stable disease (SD). Seven progressed, deemed non-responders (NR). VDAC1 expression in peripheral CD8+ T cells was higher at baseline in CBR than NR but decreased in CBR upon treatment. On-treatment peripheral CD4+ Treg levels decreased in CBR but not in NR. Analysis of tumour demonstrated PD-L1+ cells touching CD8+ T cells, and NK cells were more abundant post-treatment vs. baseline. A higher intensity of PD-L1 in tumour infiltrates at baseline, particularly in CBR vs. NR. Finally, higher levels of soluble (s)IDO, sLag3, sPD-1 observed at baseline among NR vs. CBR. CONCLUSION: Pelareorep and pembrolizumab showed modest efficacy in unselected patients, although potential immune and metabolic biomarkers were identified to warrant further evaluation.

Humoral Immunity After COVID-19 Vaccination in Chronic Lymphocytic Leukemia and Other Indolent Lymphomas: A Single-Center Observational Study.

BACKGROUND: Chronic lymphocytic leukemia (CLL) and other non-Hodgkin's lymphomas (NHLs) lead to broad immunosuppression, conferring a greater risk for morbidity and mortality from SARS-CoV-2. Our study analyzed antibody (Ab) seropositivity from SARS-CoV-2 vaccination in patients with these cancers. METHODS: In the final analysis, 240 patients were involved, and seropositivity was defined as a positive total or spike protein Ab. RESULTS: Seropositivity was 50% in CLL, 68% in WM, and 70% in the remaining NHLs. Moderna vaccination led to higher seropositivity compared to Pfizer vaccination across all cancers (64% vs. 49%; P = .022) and specifically CLL patients (59% vs. 43%; P = .029). This difference was not explainable by differences in treatment status or prior anti-CD20 monoclonal Ab therapy. In CLL patients, current or prior cancer therapy led to lower seropositivity compared to treatment-naïve patients (36% vs. 68%; P = .000019). CLL patients treated with Bruton's tyrosine kinase (BTK) inhibitors had better seropositivity after receiving the Moderna vaccination compared to Pfizer (50% vs. 23%; P = .015). Across all cancers, anti-CD20 agents within 1 year led to a lower Ab response compared to greater than one year (13% vs. 40%; P = .022), a difference which persisted after booster vaccination. CONCLUSION: Antibody response is lower in patients with indolent lymphomas compared to the general population. Lower Ab seropositivity was found in patients with a history of anti-leukemic agent therapy or those immunized with Pfizer vaccine. This data suggests that Moderna vaccination may confer a greater degree of immunity against SARS-CoV-2 in patients with indolent lymphomas.

A deep neural network framework to derive interpretable decision rules for accurate traumatic brain injury identification of infants.

OBJECTIVE: We aimed to develop a robust framework to model the complex association between clinical features and traumatic brain injury (TBI) risk in children under age two, and identify significant features to derive clinical decision rules for triage decisions. METHODS: In this retrospective study, four frequently used machine learning models, i.e., support vector machine (SVM), random forest (RF), deep neural network (DNN), and XGBoost (XGB), were compared to identify significant clinical features from 24 input features associated with the TBI risk in children under age two under the permutation feature importance test (PermFIT) framework by using the publicly available data set from the Pediatric Emergency Care Applied Research Network (PECARN) study. The prediction accuracy was determined by comparing the predicted TBI status with the computed tomography (CT) scan results since CT scan is the gold standard for diagnosing TBI. RESULTS: At a significance level of [Formula: see text], DNN, RF, XGB, and SVM identified 9, 1, 2,  and 4 significant features, respectively. In a comparison of accuracy (Accuracy), the area under the curve (AUC), and the precision-recall area under the curve (PR-AUC), the permutation feature importance test for DNN model was the most powerful framework for identifying significant features and outperformed other methods, i.e., RF, XGB, and SVM, with Accuracy, AUC, and PR-AUC as 0.915, 0.794, and 0.974, respectively. CONCLUSION: These results indicate that the PermFIT-DNN framework robustly identifies significant clinical features associated with TBI status and improves prediction performance. The findings could be used to inform the development of clinical decision tools designed to inform triage decisions.

Slow Dynamics of Acute Postoperative Pain Intensity Time Series Determined via Wavelet Analysis Are Associated With the Risk of Severe Postoperative Day 30 Pain.

BACKGROUND: Evidence suggests that increased early postoperative pain (POP) intensities are associated with increased pain in the weeks following surgery. However, it remains unclear which temporal aspects of this early POP relate to later pain experience. In this prospective cohort study, we used wavelet analysis of clinically captured POP intensity data on postoperative days 1 and 2 to characterize slow/fast dynamics of POP intensities and predict pain outcomes on postoperative day 30. METHODS: The study used clinical POP time series from the first 48 hours following surgery from 218 patients to predict their mean POP on postoperative day 30. We first used wavelet analysis to approximate the POP series and to represent the series at different time scales to characterize the early temporal profile of acute POP in the first 2 postoperative days. We then used the wavelet coefficients alongside demographic parameters as inputs to a neural network to predict the risk of severe pain 30 days after surgery. RESULTS: Slow dynamic approximation components, but not fast dynamic detailed components, were linked to pain intensity on postoperative day 30. Despite imbalanced outcome rates, using wavelet decomposition along with a neural network for classification, the model achieved an F score of 0.79 and area under the receiver operating characteristic curve of 0.74 on test-set data for classifying pain intensities on postoperative day 30. The wavelet-based approach outperformed logistic regression (F score of 0.31) and neural network (F score of 0.22) classifiers that were restricted to sociodemographic variables and linear trajectories of pain intensities. CONCLUSIONS: These findings identify latent mechanistic information within the temporal domain of clinically documented acute POP intensity ratings, which are accessible via wavelet analysis, and demonstrate that such temporal patterns inform pain outcomes at postoperative day 30.

On kernel machine learning for propensity score estimation under complex confounding structures.

Post marketing data offer rich information and cost-effective resources for physicians and policy-makers to address some critical scientific questions in clinical practice. However, the complex confounding structures (e.g., nonlinear and nonadditive interactions) embedded in these observational data often pose major analytical challenges for proper analysis to draw valid conclusions. Furthermore, often made available as electronic health records (EHRs), these data are usually massive with hundreds of thousands observational records, which introduce additional computational challenges. In this paper, for comparative effectiveness analysis, we propose a statistically robust yet computationally efficient propensity score (PS) approach to adjust for the complex confounding structures. Specifically, we propose a kernel-based machine learning method for flexibly and robustly PS modeling to obtain valid PS estimation from observational data with complex confounding structures. The estimated propensity score is then used in the second stage analysis to obtain the consistent average treatment effect estimate. An empirical variance estimator based on the bootstrap is adopted. A split-and-merge algorithm is further developed to reduce the computational workload of the proposed method for big data, and to obtain a valid variance estimator of the average treatment effect estimate as a by-product. As shown by extensive numerical studies and an application to postoperative pain EHR data comparative effectiveness analysis, the proposed approach consistently outperforms other competing methods, demonstrating its practical utility.

Bagging and deep learning in optimal individualized treatment rules.

An ENsemble Deep Learning Optimal Treatment (EndLot) approach is proposed for personalized medicine problems. The statistical framework of the proposed method is based on the outcome weighted learning (OWL) framework which transforms the optimal decision rule problem into a weighted classification problem. We further employ an ensemble of deep neural networks (DNNs) to learn the optimal decision rule. Utilizing the flexibility of DNNs and the stability of bootstrap aggregation, the proposed method achieves a considerable improvement over existing methods. An R package "ITRlearn" is developed to implement the proposed method. Numerical performance is demonstrated via simulation studies and a real data analysis of the Cancer Cell Line Encyclopedia data.

SECIMTools: a suite of metabolomics data analysis tools.

BACKGROUND: Metabolomics has the promise to transform the area of personalized medicine with the rapid development of high throughput technology for untargeted analysis of metabolites. Open access, easy to use, analytic tools that are broadly accessible to the biological community need to be developed. While technology used in metabolomics varies, most metabolomics studies have a set of features identified. Galaxy is an open access platform that enables scientists at all levels to interact with big data. Galaxy promotes reproducibility by saving histories and enabling the sharing workflows among scientists. RESULTS: SECIMTools (SouthEast Center for Integrated Metabolomics) is a set of Python applications that are available both as standalone tools and wrapped for use in Galaxy. The suite includes a comprehensive set of quality control metrics (retention time window evaluation and various peak evaluation tools), visualization techniques (hierarchical cluster heatmap, principal component analysis, modular modularity clustering), basic statistical analysis methods (partial least squares - discriminant analysis, analysis of variance, t-test, Kruskal-Wallis non-parametric test), advanced classification methods (random forest, support vector machines), and advanced variable selection tools (least absolute shrinkage and selection operator LASSO and Elastic Net). CONCLUSIONS: SECIMTools leverages the Galaxy platform and enables integrated workflows for metabolomics data analysis made from building blocks designed for easy use and interpretability. Standard data formats and a set of utilities allow arbitrary linkages between tools to encourage novel workflow designs. The Galaxy framework enables future data integration for metabolomics studies with other omics data.

Detection of clinically significant prostate cancer following initial omission of biopsy in multiparametric MRI era.

BACKGROUND: Multiparametric prostate MRI (mpMRI) is being increasingly adopted for work-up of prostate cancer. For patients selected to omit biopsy, we identified factors associated with repeat MRI, eventual prostate biopsy, and subsequent detection of clinically significant prostate cancer (csPCa, Grade Group ≥2). METHODS: We identified biopsy-naïve men presenting with PSA 2-20 ng/mL (March 2018-June 2021) undergoing initial mpMRI with PIRADS 1-3 lesions who were not selected for biopsy with ≥6 months follow-up. We examined factors associated with repeat mpMRI, progression to biopsy, and subsequent detection of csPCa with univariable and multivariable logistic regression. RESULTS: Of 1494 men, 31% (463/1494) did not pursue biopsy. PSA density (PSAD) ≤ 0.1, prostate health index (PHI) < 55, and PIRADS 1-2 were associated with omission of prostate biopsy. csPCa diagnosis-free survival was 97.6% (326/334) with median follow up of 23.1 months (IQR 15.1-34.6 months). Black race, PSA, PHI, PSA density, and PSA and PHI velocity were significant predictors of undergoing repeat mpMRI (15.6%, 52/334) and subsequent biopsy (8.4%, 28/334). 8 men were subsequently diagnosed with csPCa (N = 7 on prostate biopsy; N = 1 incidentally on holmium enucleation of prostate). All patients diagnosed with csPCa had PIRADS 4-5 on repeat mpMRI. CONCLUSIONS: The subsequent detection rate of csPCa among patients not initially biopsied after mpMRI was low at 2.4%. Decisions to omit biopsy after initial reassuring PHI, PSAD, and mpMRI appear safe with subsequent reassuring serum biomarkers and for cause mpMRI during follow-up.

Incorporating longitudinal history of risk factors into atherosclerotic cardiovascular disease risk prediction using deep learning.

It is increasingly clear that longitudinal risk factor levels and trajectories are related to risk for atherosclerotic cardiovascular disease (ASCVD) above and beyond single measures. Currently used in clinical care, the Pooled Cohort Equations (PCE) are based on regression methods that predict ASCVD risk based on cross-sectional risk factor levels. Deep learning (DL) models have been developed to incorporate longitudinal data for risk prediction but its benefit for ASCVD risk prediction relative to the traditional Pooled Cohort Equations (PCE) remain unknown. Our study included 15,565 participants from four cardiovascular disease cohorts free of baseline ASCVD who were followed for adjudicated ASCVD. Ten-year ASCVD risk was calculated in the training set using our benchmark, the PCE, and a longitudinal DL model, Dynamic-DeepHit. Predictors included those incorporated in the PCE: sex, race, age, total cholesterol, high density lipid cholesterol, systolic and diastolic blood pressure, diabetes, hypertension treatment and smoking. The discrimination and calibration performance of the two models were evaluated in an overall hold-out testing dataset. Of the 15,565 participants in our dataset, 2170 (13.9%) developed ASCVD. The performance of the longitudinal DL model that incorporated 8 years of longitudinal risk factor data improved upon that of the PCE [AUROC: 0.815 (CI 0.782-0.844) vs 0.792 (CI 0.760-0.825)] and the net reclassification index was 0.385. The brier score for the DL model was 0.0514 compared with 0.0542 in the PCE. Incorporating longitudinal risk factors in ASCVD risk prediction using DL can improve model discrimination and calibration.

A phase 1-2 trial of DA-EPOCH-R plus ixazomib for MYC-aberrant lymphoid malignancies: the DACIPHOR regimen.

ABSTRACT: MYC-aberrant non-Hodgkin lymphoma (NHL) is associated with poor outcomes with conventional chemotherapy. Ixazomib is an orally bioavailable proteasome inhibitor that targets drivers of MYC expression and has demonstrated preclinical activity in aggressive MYC-aberrant NHL. We conducted a phase 1/2 study evaluating the safety and efficacy of DA-EPOCH-R with adjunctive ixazomib in aggressive MYC-aberrant NHL. For induction, patients received 6 cycles of DA-EPOCH-R with ixazomib administered twice per 21-day cycle; responders continued weekly ixazomib maintenance for up to 1 year. Primary objectives were to determine the maximum tolerated dose in phase 1 and efficacy of DA-EPOCH-R with ixazomib as measured by 12-month progression-free survival (PFS) rate in phase 2. Thirty-six patients were evaluable for response. Median age was 63 years (range, 31-77) and 44% had double-hit lymphoma (DHL)/triple-hit lymphoma (THL). In phase 1, 3 mg of ixazomib was established as recommended phase 2 dose. Twenty-nine (76.3%) patients completed 6 cycles of DA-EPOCH-R and 25 (65.8%) underwent dose escalations. The ORR after induction was 97% (95% confidence interval, 81-100) with a CR rate of 69%. At median follow-up of 18.8 months, the 12-month PFS and overall survival (OS) rates were 78% and 86%, respectively. For DHL/THL vs dual expressor lymphomas (DEL), 12-month PFS rates were 53% vs 95% and 12-month OS rates were 65% vs 100%, respectively. Grade ≥3 toxicities were predominantly hematologic. Twenty-seven (75%) of patients experienced neuropathy, nearly all low-grade. DA-EPOCH-R induction with adjunctive ixazomib is feasible and appears effective in patients with DEL. This trial was registered at www.clinicaltrials.gov as #NCT02481310.

Age-stratified comorbid and pharmacologic analysis of patients with glioblastoma.

Background: Increased age is a strong and unfavorable prognostic factor for patients with glioblastoma (GBM). However, the relationships between stratified patient age, comorbidities, and medications have yet to be explored in GBM patient survival analyses. Objective: To evaluate co-morbid conditions, tumor-related symptoms, medication prescriptions, and subject age for patients with GBM and to establish potential targets for prospective studies. Methods: Electronic health records for 565 patients with IDHwt GBM were evaluated at a single center between January 1, 2000 and August 9, 2021 were retrospectively assessed. Data were stratified by MGMT promoter methylation status when available and were used to construct multivariable time-dependent cox models and intra-cohort hazards. Results: Younger (<65 years of age) but not older (≥65 years) GBM patients demonstrated a worse prognosis with movement related disabilities (P < 0.0001), gait/balance difficulty (P = 0.04) and weakness (P = 0.007), as well as psychiatric conditions, mental health disorders (P = 0.002) and anxiety (P = 0.001). In contrast, older but not younger GBM patients demonstrated a worse prognosis with epilepsy (P = 0.039). Both groups had worse survival with confusion/altered mental status (P = 0.023 vs < 0.000) and an improved survival with a Temozolomide prescription. Older but not younger GBM patients experienced an improved hazard with a prescription of ace-inhibitor medications (P = 0.048). Conclusion: Age-dependent novel associations between clinical symptoms and medications prescribed for co-morbid conditions were demonstrated in patients with GBM. The results of the current work support future mechanistic studies that investigate the negative relationship(s) between increased age, comorbidities, and drug therapies for differential clinical decision-making across the lifespan of patients with GBM.

An Outcomes Comparison Between Holmium Laser Enucleation of the Prostate, Open Simple Prostatectomy, and Robotic Simple Prostatectomy for Large Gland Benign Prostatic Hypertrophy.

OBJECTIVE: To compare perioperative outcomes between Holmium laser enucleation of the prostate (HoLEP), open simple prostatectomy (OSP), and robotic simple prostatectomy (RSP) for large prostates (> 80 cc). MATERIALS AND METHODS: A retrospective study of 340 patients who underwent HoLEP (n = 209), OSP (n = 66), or RSP (n = 65) at a large academic medical center between January 2013 - September 2021 was performed. Length of stay (LOS), operative time, catheter duration, estimated blood loss (EBL), blood transfusion, and 30-day ED visits and readmissions were compared between the three groups. Univariate analyses consisted of ANOVA with Tukey's corrections and Chi-square tests. Linear and multivariate logistic regression was also performed. All tests were two-sided and a p-value <0.05 was pre-determined to be statistically significant. Analyses were performed with SAS v9.4. RESULTS: HoLEP was found to have the shortest: operative time (1.4 vs 2.7 vs 3.8h), LOS (0.65 vs 4.2 vs 2.6d), and catheter duration (0.38 vs 9.9 vs 11.2d) compared to OSP and RSP, respectively (all P <.0001). HoLEP also had the lowest EBL (66 vs 795 vs 326 mL, P <.0001). HoLEP and RSP had a lower risk of blood transfusion compared to OSP (P <.0001). These associations remained significant on multivariable analyses. CONCLUSION: HoLEP is a minimally invasive treatment option for large prostates that was found to have shorter operative time, LOS, and catheter duration as well as lower EBL compared to OSP and RSP.

Phase I study of novel SYK inhibitor TAK-659 (mivavotinib) in combination with R-CHOP for front-line treatment of high-risk diffuse large B-cell lymphoma.

Background: TAK-659, a novel oral SYK inhibitor, has demonstrated efficacy in heavily pretreated diffuse large B-cell lymphoma (DLBCL). We report results of a phase I single-institution escalation study of front-line treatment with R-CHOP and TAK-659 in treatment-naïve high-risk DLBCL. Methods: Patients with high-risk DLBCL were treated with R-CHOP for 1 cycle, followed by combined R-CHOP and TAK-659 for an additional five cycles, with TAK-659 dosing escalated from 60 mg, to 80 mg, to 100 mg daily, based on a 3 + 3 design. The primary objective was to determine the safety and establish the maximum tolerated dose (MTD) of TAK-659 in this setting. Results: Twelve patients were enrolled. Dose level 3 (100 mg) was established as the MTD. Dose level 1 (60 mg) maintained a similar area under the curve (AUC) to the MTD. With a median follow-up of 21 months, 92% of patients achieved complete response (CR). The most common treatment-emergent adverse events were lymphopenia (100%), infection (50%, n = 3 opportunistic), aspartate aminotransferase elevation (100%), and alanine aminotransferase elevation (83%). Conclusion: A TAK-659 dose of 60 mg was well tolerated, did not require dose modifications, and maintained a similar AUC to the MTD. The combination of R-CHOP and TAK-659 in patients with newly diagnosed high-risk DLBCL produces promising CR rates.

Real-world use of MRI for risk stratification prior to prostate biopsy.

BACKGROUND: The utilization of MRI to risk stratify elevated PSA prior to prostate biopsy has been inconsistently adopted and varies considerably by practice setting. This study aims to evaluate the usage and performance of MRI as an advanced risk stratification tool of elevated PSA prior to biopsy and identify factors associated with differential utilization of MRI at a large academic setting with ready access to 3T multiparametric MRI of the prostate. METHODS: A retrospective single-center study of 2900 men presenting with elevated PSA 2-20 ng/mL from 2018 through 2021 was conducted. We analyzed trends in MRI utilization and outcomes of prostate biopsy by MRI usage. Univariate and multivariate logistic regressions were performed to calculate odds ratios to identify patient- and provider-level predictors of MRI usage. RESULTS: Rates of prebiopsy MRI utilization increased from 56% in 2018 to 89% in 2021 (p < 0.001). Prebiopsy MRI led to biopsy avoidance in 31% of men. MRI usage enhanced detection of clinically significant prostate cancer by 13% and reduced identification of Gleason Grade Group 1 disease by 3% and negative biopsies by 10% (p < 0.001). Men who received MRI were more likely to be younger than 75 years in age and have private or Medicare insurance, PSA >4 ng/mL, and PHI >27. In both univariate and multivariate analysis, black race and Medicaid insurance were associated with reduced MRI utilization (all p < 0.001). Urologic provider was an independent predictor of MRI usage (p < 0.001). CONCLUSIONS: Use of MRI as a risk stratification tool for elevated PSA rose during this 4-year study period. Men who self-identify as black or men with Medicaid coverage have diminished rates of MRI usage. Considerable provider-level variability in MRI use was observed. Future research aimed at identifying factors affecting implementation of MRI as a routine risk assessment tool is warranted.

Resource and Reimbursement Barriers to Comprehensive Cancer Care Delivery: An Analysis of Association of Community Cancer Centers Survey Data.

PURPOSE: Comprehensive cancer care (CCC) delivery is recommended in guidelines and considered essential for high-quality cancer management. Barriers, such as insufficient reimbursement, prevent consistent access to and delivery of CCC. Association of Community Cancer Centers conducted a national survey to elucidate capacity and barriers to CCC delivery to inform policy and value-based payment reform. METHODS: Survey methodology included item generation with expert review, iterative piloting, and cognitive validity testing. In the final instrument, 27 supportive oncology services were assessed for availability, reasons not offered, and coverage/reimbursement. RESULTS: 204 of 704 member programs completed survey questions. Despite most services being reported as offered, a minority were funded through insurance reimbursement. The services least likely to obtain reimbursement were those that address practical and family/childcare needs (0.7%), caregiver support (1.5%), advanced care directives (1.7%), spiritual services (1.8%), and navigation (2.7%). These findings did not vary by region or practice type. CONCLUSION: There is a lack of sufficient reimbursement, staffing, and budget to provide CCC across the United States. Care models and reimbursement policies must include CCC services to optimize delivery of cancer care.

Multicenter phase 2 study of romidepsin plus lenalidomide for previously untreated peripheral T-cell lymphoma.

Peripheral T-cell lymphomas (PTCLs) are associated with poor prognosis when treated with cytotoxic chemotherapy. We report the findings of a phase 2 study evaluating a chemotherapy-free combination of romidepsin plus lenalidomide as initial treatment for patients with PTCL who were aged >60 years or noncandidates for chemotherapy. Treatment was initiated with romidepsin 10 mg/m2 IV on days 1, 8, and 15 and lenalidomide 25 mg taken orally from days 1 to 21 of 28-day cycle for up to 1 year. The primary objective was overall response rate (ORR). Secondary objectives included safety and survival. The study enrolled 29 patients with a median age of 75 years, including 16 (55%) angioimmunoblastic T-cell lymphoma (AITL), 10 (34%) PTCL- not otherwise specified, 2 ATLL, and 1 EATL. Grade 3 to 4 hematologic toxicities included neutropenia (45%), thrombocytopenia (34%), and anemia (28%). Grade 3 to 4 nonhematologic toxicities included hyponatremia (45%), hypertension (38%), hypoalbuminemia (24%), fatigue (17%), hyperglycemia (14%), hypokalemia (14%), dehydration (10%), and infection (10%). At median follow-up of 15.7 months, 23 patients were evaluable and received a median treatment of 6 cycles. The ORR was 65.2% with complete response (CR) at 26.1%, including 78.6% ORR and 35.7% CR for AITL. Median duration of response was 10.7 months, with 27.1 months for patients achieving CR. The estimated 2-year progression-free survival was 31.5%, and 2-year overall survival was 49.5%. This study provides the first demonstration that the biologic combination of romidepsin and lenalidomide is feasible and effective as initial therapy for PTCL and warrants further evaluation. This trial was registered at www.clinicaltrials.gov as #NCT02232516.

Ibrutinib maintenance after frontline treatment in patients with mantle cell lymphoma.

Maintenance rituximab in mantle cell lymphoma (MCL) has improved survival and supports exploration of maintenance with novel agents. We evaluated the safety and efficacy of ibrutinib maintenance (I-M) after induction in patients with treatment-naive MCL. Patients with MCL with complete response (CR) or partial response to frontline chemoimmunotherapy ± autologous stem cell transplantation (auto-SCT) received I-M 560 mg daily for up to 4 years. Primary objective was 3-year progression-free survival (PFS) rate from initiation of I-M. Minimal residual disease (MRD) assessments by next-generation sequencing (NGS) on peripheral blood were measured before I-M initiation and at 1, 6, and 18 to 24 months after initiation. Among 36 patients, the median age was 60 years (range, 46-90). For frontline treatment, 18 patients (50%) had consolidation with auto-SCT in CR1 before I-M. At median follow-up of 55.7 months, 17 patients (47%) completed full course I-M (median, 37.5 cycles; range, 2-52). The 3-year PFS and overall survival (OS) rates were 94% and 97%, respectively. With prior auto-SCT, 3-year PFS and OS rates were both 100%. The most common treatment-related adverse event with I-M was infection (n = 31; 86%), typically low grade; the most common grade 3/4 toxicities were hematologic. In 22 patients with MRD assessments, all were MRD negative after induction. Six became MRD positive on I-M, with 2 reverting to MRD-negative status with continued I-M, and all maintain radiographic CR with the exception of 1 with disease progression. I-M is feasible in MCL after frontline chemoimmunotherapy with manageable toxicities although significant. Changes in NGS-MRD were noted in limited patients during maintenance with few progression and survival events. This trial was registered at www.clinicaltrials.gov as #NCT02242097.

Incorporating longitudinal history of risk factors into atherosclerotic cardiovascular disease risk prediction using deep learning.

Background: It is increasingly clear that longitudinal risk factor levels and trajectories are related to risk for atherosclerotic cardiovascular disease (ASCVD) above and beyond single measures. Currently used in clinical care, the Pooled Cohort Equations (PCE) are based on regression methods that predict ASCVD risk based on cross-sectional risk factor levels. Deep learning (DL) models have been developed to incorporate longitudinal data for risk prediction but its benefit for ASCVD risk prediction relative to the traditional Pooled Cohort Equations (PCE) remain unknown. Objective: To develop a ASCVD risk prediction model that incorporates longitudinal risk factors using deep learning. Methods: Our study included 15,565 participants from four cardiovascular disease cohorts free of baseline ASCVD who were followed for adjudicated ASCVD. Ten-year ASCVD risk was calculated in the training set using our benchmark, the PCE, and a longitudinal DL model, Dynamic-DeepHit. Predictors included those incorporated in the PCE: sex, race, age, total cholesterol, high density lipid cholesterol, systolic and diastolic blood pressure, diabetes, hypertension treatment and smoking. The discrimination and calibration performance of the two models were evaluated in an overall hold-out testing dataset. Results: Of the 15,565 participants in our dataset, 2,170 (13.9%) developed ASCVD. The performance of the longitudinal DL model that incorporated 8 years of longitudinal risk factor data improved upon that of the PCE [AUROC: 0.815 (CI: 0.782-0.844) vs 0.792 (CI: 0.760-0.825)] and the net reclassification index was 0.385. The brier score for the DL model was 0.0514 compared with 0.0542 in the PCE. Conclusion: Incorporating longitudinal risk factors in ASCVD risk prediction using DL can improve model discrimination and calibration.