TP53-mutant mantle cell lymphoma (MCL) is associated with poor survival outcomes with standard chemoimmunotherapy. We conducted a multicenter, phase 2 study of zanubrutinib, obinutuzumab, and venetoclax (BOVen) in untreated patients with MCL with a TP53 mutation. Patients initially received 160 mg zanubrutinib twice daily and obinutuzumab. Obinutuzumab at a dose of 1000 mg was given on cycle 1 day 1, 8, and 15, and on day 1 of cycles 2 to 8. After 2 cycles, venetoclax was added with weekly dose ramp-up to 400 mg daily. After 24 cycles, if patients were in complete remission with undetectable minimal residual disease (uMRD) using an immunosequencing assay, treatment was discontinued. The primary end point was met if ≥11 patients were progression free at 2 years. The study included 25 patients with untreated MCL with a TP53 mutation. The best overall response rate was 96% (24/25) and the complete response rate was 88% (22/25). Frequency of uMRD at a sensitivity level of 1 × 10–5 and uMRD at a sensitivity level of 1 × 10–6 at cycle 13 was 95% (18/19) and 84% (16/19), respectively. With a median follow-up of 28.2 months, the 2-year progression-free, disease-specific, and overall survival were 72%, 91%, and 76%, respectively. Common side effects were generally low grade and included diarrhea (64%), neutropenia (32%), and infusion-related reactions (24%). BOVen was well tolerated and met its primary efficacy end point in TP53-mutant MCL. These data support its use and ongoing evaluation. This trial was registered at www.ClinicalTrials.gov as #NCT03824483.

Pancreatic adenocarcinoma (PDAC) is the most common pancreatic cancer and is associated with poor prognosis, a high mortality rate, and a substantial number of healthy life years lost. Surgical resection is the primary treatment option for patients with resectable disease; however, only 10–20% of all patients with PDAC are eligible for resection at the time of diagnosis. In this context, neoadjuvant therapy has the potential to increase the number of patients who are eligible for resection, thereby improving the overall survival rate. For patients who undergo neoadjuvant therapy, computed tomography (CT) remains the primary imaging tool for assessing treatment response. Nevertheless, the interpretation of imaging findings in this context remains challenging, given the similarity between viable tumor and treatment-related changes following neoadjuvant therapy. In this review, following an overview of the various treatment options for PDAC according to its resectability status, we will describe the key challenges regarding CT-based evaluation of PDAC treatment response following neoadjuvant therapy, as well as summarize the literature on CT-based evaluation of PDAC treatment response, including the use of radiomics. Finally, we will outline key recommendations for the management of PDAC after neoadjuvant therapy, taking into consideration CT-based findings.

Generating Real World Evidence (RWE) on disease responses from radiological reports is important for understanding cancer treatment effectiveness and developing personalized treatment. A lack of standardization in reporting among radiologists impacts the feasibility of large-scale interpretation of disease response. This study examines the utility of applying natural language processing (NLP) to the large-scale interpretation of disease responses using a standardized oncologic response lexicon (OR-RADS) to facilitate RWE collection. Radiologists annotated 3503 retrospectively collected clinical impressions from radiological reports across several cancer types with one of seven OR-RADS categories. A Bidirectional Encoder Representations from Transformers (BERT) model was trained on this dataset with an 80–20% train/test split to perform multiclass and single-class classification tasks using the OR-RADS. Radiologists also performed the classification to compare human and model performance. The model achieved accuracies from 95 to 99% across all classification tasks, performing better in single-class tasks compared to the multiclass task and producing minimal misclassifications, which pertained mostly to overpredicting the equivocal and mixed OR-RADS labels. Human accuracy ranged from 74 to 93% across all classification tasks, performing better on single-class tasks. This study demonstrates the feasibility of the BERT NLP model in predicting disease response in cancer patients, exceeding human performance, and encourages the use of the standardized OR-RADS lexicon to improve large-scale prediction accuracy.