Sovanlal Mukherjee: Excited to Share New AI Advances in Pancreatic Cancer Imaging
Sovanlal Mukherjee/ mayo.edu

Sovanlal Mukherjee: Excited to Share New AI Advances in Pancreatic Cancer Imaging

Sovanlal Mukherjee, Senior Scientist at Mayo Clinic, shared a post on LinkedIn:

“I’m excited to share our latest publication in npj Precision Oncology (Nature Portfolio).

Pancreatic ductal adenocarcinoma (PDAC) remains one of the most challenging cancers to assess accurately on CT. Reliable volumetric tumor segmentation is essential for quantitative imaging, radiomics, treatment response assessment, and precision oncology – but developing models that are robust enough for clinical translation has remained difficult.

In this work, we developed and validated an anatomically constrained deep learning framework using 1,859 biopsy-confirmed, treatment-naïve PDAC CT examinations. The model demonstrated:

  •  Strong and consistent performance across internal and external validation cohorts.
  •  Robustness across scanner vendors, institutions, acquisition protocols, and imaging eras.
  •  Superior performance compared with a state-of-the-art 3D vision transformer under matched training conditions.
  •  Performance approaching expert-reader variability on challenging cases, supporting its potential as a reproducible foundation for volumetric tumor quantification.

Our hope is that this work will help enable more reliable imaging biomarkers for pancreatic cancer, supporting future applications in treatment response assessment, radiomics, and multimodal AI models that integrate imaging with clinical and molecular data.

This project was the result of an incredible multidisciplinary collaboration. I’m deeply grateful to all my co-authors, collaborators who made this research possible.

I look forward to seeing how the field continues to advance AI toward clinically meaningful applications in pancreatic cancer.

Title: Anatomically constrained deep learning for clinical-grade volumetric pancreatic cancer segmentation: development, validation, and architectural benchmarking

Authors: Sovanlal Mukherjee, Khurram Khaliq Bhinder, Armin Zarrintan, Ajith Antony, Angela Ammirabile, Ahmed Jadoon, Subhosree Dey, Baloy J. Talukdar, Takeru Yamaguchi, Rondell P. Graham, Suresh T. Chari, Ajit Harishkumar Goenka

Read The Full Article.

Sovanlal Mukherjee: Excited to Share New AI Advances in Pancreatic Cancer Imaging

Other articles about Pancreatic Cancer on OncoDaily.