Pathology companies that are developing artificial intelligence algorithms on whole slide images from patient biopsies are popping up everywhere and all the time. These companies state that they want to improve healthcare by improving pathology diagnoses. Since the number of pathologists throughout the world is insufficient and the number of pathology biopsies and resections is increasing, Artificial Intelligence is one way to automate the tedious and boring aspects of pathology to free up pathologists’ time for more difficult interpretations. The companies claim that AI will improve the detection of disease and the standardization of tumor grading. These companies range from well-funded startups to giant internet companies, such as Google and Baidu. There are over 30 companies around the world that profess to be performing artificial intelligence on digital pathology. Almost half of these companies are in the USA and Canada, Twenty-three percent are in Europe (UK, Ireland, Germany, Denmark, Sweden Norway and Finland). Nineteen percent are in Asia (Korea, India, China, Taiwan) and ten percent are in Israel
I admit that I am fascinated by this sudden development. I intend to profile many of these companies in my blog. Some companies will be winners. Some will be acquired by much larger companies like Philips, Leica, and Roche who want to complete their portfolio of pathology products, as well as, participate in this potentially game changing area of diagnostics. Here are some companies to watch .
Paige is a New York City based company that is a spinoff from Memorial Sloan Kettering Cancer Center. Thomas Fuchs, Dr. Sc is the Founder and Chief Science Officer. Paige is an acronym for Pathology Artificial Intelligence Guidance Engine. Paige has an exclusive license with MSKCC for access to its pathology department’s 25 million slides. MSKCC is apparently digitizing over a 1000 slides a day. Paige uses machine learning algorithms with Convolutional Neural Networks. Paige has secured over $25 million in Series A financing from Beyer Capital
PathAI is a Cambridge, MA based company that is a spinoff from Harvard Medical School/Beth Israel Deaconess Medical Center. Andrew Beck MD PhD, anatomic pathologist and PhD in informatics from Stanford cofounded PathAI with Aditya Khosla PhD, computer scientist from MIT and Stanford. PathAI also uses machine learning with CNN to develop diagnostic algorithms. Members of PathAI have participated in Camelyon16 and did very well in the competition. PathAI has a partnership with Novartis. I would expect PathAI to leverage its connections to Harvard’s Gordon Center for Medical Imaging, Beth Israel Deaconess and MIT. PathAI has recently received over $11 M in series A funding.
Proscia is an digital pathology image workflow, image management system and pathology AI company based in Philadelphia, PA. Cofounders David West, CEO; Coleman Stavich, CTO and Nathan Buchbinder, Chief Product Officer are biomedical engineers and computer scientists from Johns Hopkins University, Moffett Cancer Center and University of Pittsburgh. Advanced Pathology Associates of Maryland have adopted Proscia’s digital workflow platform. Proscia is refining and hosting an oral pathology artificial intelligence predictive algorithm to predict the risk of developing oral squamous cell cancer from oral biopsies for Canadian company ProteocyteAI. Proscia has also signed an agreement with Dermpath Lab of Central States (DLCS) to develop diagnostic AI tools to identify the common and typical skin biopsy diagnoses, in order to free up dermatopathologist time for more difficult diagnoses. Proscia has $ 8.3 million in Series A financing.
Google AI Healthcare Google’s parent company, Alphabet, is reorganizing its healthcare investments, research and subsidiaries under Google Health. Artificial intelligence is and will continue to be big part of Google/Alphabet portfolio. Already, Google has developed a deep learning algorithm to identify diabetic retinopathy from retinal fundus images. In pathology, Google participated in Camelyon16 competition to identify metastases in sentinel lymph nodes in patients with breast cancer, eventually achieving an Area Under the Curve (AUC) of 97%. Recently Google AI participated in a deep learning study in prostate cancer. The deep learning program had greater accuracy in Gleason grading than generalist pathologist. Google’s business model in pathology is not entirely clear. Google maybe looking to foster digital pathology and AI in pathology because the storage and analysis of large image files will be beneficial to Google’s Cloud services. Jason Hipp, MD PhD is a pathologist in Google AI’s pathology division.
Indica Labs is based in Corrales NM is an digital pathology image analysis company. It has multiple applications in quantitative image analysis including multiplex IHC, especially useful for immunotherapy predictive analysis, branded as HALO IMAGE ANALYSIS. Indica has an AI offering called HALO AI. It is a deep learning algorithm that pathologist can use remotely. This algorithm performed very well in the Camelyon17 Challenge to identify metastatic breast cancer in sentinel lymph nodes and to automatically provide the lymph node stage in TNM. Indica Lab also offers an image management system. Its CEO, Steven Hashagen was previously a software engineer at Aperio and Kate Tunstall, PhD, its Chief Science Officer was previously at Aperio and Biocyte. Indica Labs recently received $ 3 million from a New Mexico economic development fund.
In my next blog post, I will profile five more companies developing pathology AI.