LungLife is an American diagnostics company focused on using technology for early lung cancer detection. It provides certainty in the early-stage detection of lung cancer. The company is developing a series of companion diagnostics for later-stage treatments while anticipating patients’ probable therapeutic responses.
The Challenge
Analyzing massive volumes of data of up to 15,000 microscopic images per patient was a substantial challenge for LungLife. Manual analysis was time-consuming, leading to delayed cancer detection, higher false positives, and missed circulating tumor cell diagnostics.
LungLife wanted to leverage AI and machine learning (ML) to overcome these challenges. The company was looking to significantly impact patient outcomes and clinical decision-making by detecting lung cancer in the early stage.
The Solution
To develop AI and ML algorithms for LungLife’s diagnostic technology, Persistent developed a deep learning-based segmentation model. We deployed annotation tools to ensure enhanced model accuracy of microscopic images. We also developed UI-based solutions to help LungLife efficiently verify cells and classify them into circulating tumor cells, single gain cells, single deletion cells, and normal cells.
The Outcome
LungLife utilized the new solution to significantly reduce diagnosis time by about 70% . With 62% less false positive cases, the accuracy rate also dramatically improved. These results accelerated LungLife’s efforts to greatly reduce the impact of a disease that claims approx. 400 lives per day.
Technology Used
- AWS
- AngularJS
- Keras
- TensorFlow
