As one of the most ubiquitous diagnostic imaging tests in medical practice, chest radiography requires timely reporting of potential findings and diagnosis of diseases in the images. Automated, fast, and reliable detection of diseases based on chest radiography is a critical step in radiology workflow. In this work, we developed and evaluated various deep convolutional neural networks (CNN) for differentiating between normal and abnormal frontal chest radiographs, in order to help alert radiologists and clinicians of potential abnormal findings as a means of work list triaging and reporting prioritization. A CNN-based model achieved an AUC of 0.9824 ± 0.0043 (with an accuracy of 94.64 ± 0.45%, a sensitivity of 96.50 ± 0.36% and a specificity of 92.86 ± 0.48%) for normal versus abnormal chest radiograph classification. The CNN model obtained an AUC of 0.9804 ± 0.0032 (with an accuracy of 94.71 ± 0.32%, a sensitivity of 92.20 ± 0.34% and a specificity of 96.34 ± 0.31%) for normal versus lung opacity classification. Classification performance on the external dataset showed that the CNN model is likely to be highly generalizable, with an AUC of 0.9444 ± 0.0029. The CNN model pre-trained on cohorts of adult patients and fine-tuned on pediatric patients achieved an AUC of 0.9851 ± 0.0046 for normal versus pneumonia classification. Pretraining with natural images demonstrates benefit for a moderate-sized training image set of about 8500 images. The remarkable performance in diagnostic accuracy observed in this study shows that deep CNNs can accurately and effectively differentiate normal and abnormal chest radiographs, thereby providing potential benefits to radiology workflow and patient care.

The effectiveness of immune checkpoint inhibitors (ICI) is limited in pancreatic ductal adenocarcinoma (PDAC). We conducted a phase I study to evaluate the safety of ICI with stereotactic body radiation therapy (SBRT) in patients with metastatic PDAC.Patients and Methods: Patients enrolled must have received at least one line of prior systemic chemotherapy for metastatic disease. Cohorts A1 and A2 received durvalumab every 2 weeks plus either 8 Gy in one fraction of SBRT on day 1 or 25 Gy in five fractions on day À3 to þ1. Cohorts B1 and B2 received durvalumab plus tremelimumab every 4 weeks and either 8 Gy in one fraction of SBRT on day 1 or 25 Gy in five fractions on day À3 to þ1. ICIs were continued until unacceptable toxicity or disease progression. The primary objective was the safety and feasibility of treatment. Objective response was assessed in lesions not subjected to SBRT.Results: Fifty-nine patients were enrolled and 39 were evaluable for efficacy. No dose-limiting toxicities were seen. The most common adverse event was lymphopenia. Two patients achieved a partial response (one confirmed and the other unconfirmed). The overall response rate was 5.1%. Median PFS and OS was 1.7 months [95% confidence intervals (CI), 0.8-2.0 months] and 3.3 months (95% CI, 1.2-6.6 months) in cohort A1; 2.5 months (95% CI, 0.1-3.7 months) and 9.0 months (95% CI, 0.5-18.4 months) in A2; 0.9 months (95% CI, 0.7-2.1 months) and 2.1 months (95% CI, 1.1-4.3 months) in B1; and 2.3 months (95% CI, 1.9-3.4 months) and 4.2 months (95% CI, 2.9-9.3 months) in B2.Conclusions: The combination of ICI and SBRT has an acceptable safety profile and demonstrates a modest treatment benefit in patients with metastatic PDAC.

No abstract

Purpose: This study investigated the efficacy and tolerability of cabozantinib plus nivolumab (CaboNivo) in patients with metastatic urothelial carcinoma (mUC) that progressed on checkpoint inhibition (CPI). Patients and Methods: A phase I expansion cohort of patients with mUC who received prior CPI was treated with cabozantinib 40 mg/day and nivolumab 3 mg/kg every 2 weeks until disease progression/unacceptable toxicity. The primary goal was objective response rate (ORR) per RECIST v.1.1. Secondary objectives included progression-free survival (PFS), duration of response (DoR), overall survival (OS), safety, and tolerability. Results: Twenty-nine out of 30 patients enrolled were evaluable for efficacy. Median follow-up was 22.2 months. Most patients (86.7%) received prior chemotherapy and all patients received prior CPI (median seven cycles). ORR was 16.0%, with one complete response and three partial responses (PR). Among 4 responders, 2 were primary refractory, 1 had a PR, and 1 had stable disease on prior CPI. Median DoR was 33.5 months [95% confidence interval (CI), 3.7–33.5], median PFS was 3.6 months (95% CI, 2.1–5.5), and median OS was 10.4 months (95% CI, 5.8–19.5). CaboNivo decreased immunosuppressive subsets such as regulatory T cells (Tregs) and increased potential antitumor immune subsets such as nonclassical monocytes and effector T cells. A lower percentage of monocytic myeloid-derived suppressor cells (M-MDSC) and polymorphonuclear MDSCs, lower CTLA-4 and TIM-3 expression on Tregs, and higher effector CD4+ T cells at baseline were associated with better PFS and/or OS. Conclusions: CaboNivo was clinically active, well tolerated, and favorably modulated peripheral blood immune subsets in patients with mUC refractory to CPI.

Background We conducted a phase II study of the combination of pembrolizumab with capecitabine and oxaliplatin (CAPOX) in patients with advanced biliary tract carcinoma (BTC) to assess response rate and clinical efficacy. Exploratory objectives included correlative studies of immune marker expression, tumor evolution, and immune infiltration in response to treatment. Patients and Methods Adult patients with histologically confirmed BTC were enrolled and received oxaliplatin and pembrolizumab on day 1 of cycles 1-6. Capecitabine was administered orally twice daily as intermittent treatment, with the first dose on day 1 and the last dose on day 14 of cycles 1-6. Starting on cycle 7, pembrolizumab monotherapy was continued until disease progression. The primary endpoint was progression-free survival (PFS). Secondary endpoints were safety, tolerability, feasibility, and response rate. Immunohistochemistry (IHC) for PD-L1 and immune infiltrates was analyzed in paired tumor biopsies, as well as bulk transcriptome and exome profiling for five patients and single-cell RNA sequencing for one partial responder. Results Eleven patients enrolled, three of whom had received no prior systemic therapy. Treatment was well tolerated, and the most common treatment-related grade 3 or 4 adverse events were lymphocytopenia, anemia, and decreased platelet count. Three patients (27.3%) achieved a partial response, and six (54%) had stable disease. The disease control rate was 81.8%. The median PFS was 4.1 months with a 6-month PFS rate of 45.5%. Molecular profiling suggests qualitative differences in immune infiltration and clonal evolution based on response. Conclusion Capecitabine and oxaliplatin in combination with pembrolizumab is tolerable and a potentially effective treatment for refractory advanced BTC. This study highlights a design framework for the precise characterization of individual BTC tumors. Trial Registration This study was registered in ClinicalTrials.gov (NCT03111732).