Skip to content

BAMF Health Research

BAMF Health Original Research

Evaluating the Efficacy and Safety of Pluvicto in Chemotherapy-Ineligible Nonagenarians: A Descriptive Case Series The study explores the safety and efficacy of radioligand therapy for men 90+ years old with metastatic castration-resistant prostate cancer.

Men who are 90+ years old with metastatic castration-resistant (mCRPC) prostate cancer are at risk of significant side effects if they undergo chemotherapy treatment. Targeted radioligand (radiopharmaceutical) therapy presents a promising alternative. Our team evaluated the response, side effects, and quality of life of patients 90+ years old who received radioligand therapy at BAMF Health.

Study Authors

First-in-Human Total-Body PET/CT Imaging Using 89Zr-Labeled MUC5AC Antibody In A Patient With Pancreatic Adenocarcinoma The first-in-human study of a novel PET tracer for the detection of pancreatic cancer.

BAMF Health, in partnership with Nihon Medi-Physics Co., Ltd. (NMP), imaged the first patient in the world at BAMF Health using the novel molecular imaging agent NMK89 developed by NMP. This was an important first step in developing improved imaging and a potential therapy for pancreatic cancer.

Read BAMF Health’s announcement.

Study Authors

  • Harshad Kulkarni, MD
  • Kevin Maupin, PhD
  • Tina Brennan
  • Jens Forsberg, PhD
  • Dan Rogers
  • Mark Olson
  • Brandon Mancini, MD
  • Anthony Chang, PhD
  • Sreenivasa Chandana, MD, PhD
  • Ryohei Kobayashi
AI-Generated Annotations Dataset For Diverse Cancer Radiology Collections In NCI Image Data Commons BAMF Health AI team partners with National Cancer Institute on pivotal research for automatic labeling of medical images.

BAMF Health was chosen to partner with the National Cancer Institute to develop AI models to rapidly analyze scans and correctly identify cancer.

In part one of the project, BAMF analyzed approximately 3,000 lung, liver, and prostate cancer PET/CT scans and developed AI algorithms to distinguish tumors from surrounding healthy tissue. Our team then partnered with radiologists to continually adjust the AI models to improve accuracy. In part two, we followed a similar process for breast and brain cancer images. In all, our team of seven created 11 AI algorithms and analyzed more than 6,000 scans. Our results were compiled on a public dashboard and are now used by cancer experts across the globe.

Study Authors

  • Gowtham Murugesan, PhD
  • Diana McCrumb
  • Mariam Aboian
  • Tej Verma
  • Rahul Soni
  • Fatima Memon
  • Keyvan Farahani
  • Linmin Pei
  • Ulrike Wagner
  • Andrey Fedorov
  • David Clunie
  • Stephen Moore, PhD
  • Jeff VanOss
Improving Lesion Segmentation In FDG-18 Whole-Body PET/CT Scans Using Multilabel Approach: AutoPET II Challenge BAMF Health AI team competes in global competition to properly label lesions in PET/CT images using AI.

The BAMF Health AI team developed a new method to improve how artificial intelligence identifies cancerous lesions in PET/CT scans. Normally, AI models can mistake certain organs, like the liver, brain, and bladder, for cancer because they also absorb a lot of the imaging tracer. To fix this, the team trained the AI to recognize both organs and cancerous lesions separately, making its predictions more accurate. They tested their approach on a large dataset of over 1,000 patients and found it worked better than other methods. Their model ranked highest in a competitive test, showing that this technique could help doctors more accurately assess cancer and improve patient care.

Study Authors

  • Gowtham Murugesan, PhD
  • Diana McCrumb
  • Eric Brunner, PhD
  • Jithendra Kumar
  • Rahul Soni
  • Vasily Grigorash
  • Stephen Moore, PhD
  • Jeff VanOss
Head and Neck Primary Tumor Segmentation Using Deep Neural Networks And Adaptive Ensembling BAMF Health AI team competes in global competition to automatically identify head and neck cancer tumors.

The BAMF Health AI team developed an advanced artificial intelligence (AI) model to help detect and outline head and neck tumors in PET/CT scans. In medical imaging, accurately identifying tumors can be challenging because they often blend in with surrounding tissues. To improve accuracy, the team used a powerful deep learning model trained on a large dataset from the HECKTOR 2021 challenge. Their AI system learned to distinguish tumors more effectively, making it easier for doctors to assess cancer and plan treatments. Their approach performed well in the competition, showing promise for improving cancer diagnosis and patient care.

Study Authors

  • Gowtham Murugesan, PhD
  • Eric Brunner, PhD
  • Diana McCrumb
  • Jithendra Kumar
  • Jeff VanOss
  • Stephen Moore, PhD
  • Anderson Peck
  • Anthony Chang, PhD