PathAI Announces Upcoming Presentations at the 2022 American Association for Cancer Research (AACR) Annual Meeting
BOSTON—April 8, 2022— PathAI, a global leader in artificial intelligence (AI)-powered technology for pathology, today announced that their recent research will be presented at the 2022 AACR annual meeting, which will be held in New Orleans from April 8 to April 13, 2022. PathAI will share a total of six posters, three of which were developed in collaboration with pharmaceutical partners. These new findings have promising implications for improving the diagnosis and treatment of multiple cancer subtypes with the use of AI-powered digital pathology.
“Our research demonstrates how PathAI’s algorithms can uncover novel insights or generate important molecular inferences from standard pathology samples that could not be identified using traditional manually-interpreted pathology or with more expensive and less accessible molecular approaches”, said Mike Montalto, Chief Scientific Officer at PathAI. “These translational insights are critically important to help accelerate drug development and get life-saving medicines to patients more quickly.”
A key example of this capability is highlighted in the poster, “AI-powered segmentation and analysis of nuclei morphology predicts genomic and clinical markers in multiple cancer types,” which describes PathAI’s machine learning-based nuclear segmentation model that associates features of nuclear morphology with clinically-relevant molecular and genomic markers across multiple tumor types.
The model was trained using over 29,000 annotations across multiple cancer types as well as five non-cancer tissues, to identify nuclei in any tissue type. Features related to nuclear shape, texture, and color were automatically extracted from the model and evaluated for their ability to predict genomic and molecular markers important for treatment selection and disease prognosis. When applied to breast, lung, and prostate tumor tissue, these models identified key combinations of nuclear features that were predictive of biomarkers including whole genome doubling, homologous recombination deficiency, HER2 positivity or elevated Gleason grade. These models could provide a way to rapidly genotype patients for selection of the most appropriate treatments without the need for extra tissue samples or sequencing protocols.
The full list of PathAI’s poster presentations is highlighted below. More information on each abstract can be found here.
Title: Quantification of TGFβ protein levels and digital pathology-based immune phenotyping reveal biomarkers for TGF-β blockade therapy patient selection in NSCLC
Session Date and Time: Online Only. E-poster available April 8, 2022, 12:00 PM CT.
Abstract: 5099
Developed in partnership with Sanofi
Title: Machine learning models identify histological features that can predict KEAP1 mutations in lung adenocarcinoma
Session Date and Time: April 10, 2022, 1:30 PM – 5:00 PM CT
Abstract: 449
Title: AI-powered segmentation and analysis of nuclei morphology predicts genomic and clinical markers in multiple cancer types
Session Date and Time: April 10, 2022, 1:30 PM – 5:00 PM CT
Abstract: 464
Title: AIM PD-L1-NSCLC: Artificial intelligence-powered PD-L1 quantification for accurate prediction of tumor proportion score in diverse, multi-stain clinical tissue samples
Session Date and Time: April 10, 2022, 1:30 PM – 5:00 PM CT
Abstract: 471
Title: AI-powered and manual assessment of PD-L1 are comparable in predicting response to neoadjuvant atezolizumab in patients (pts) with resectable non-squamous, non-small cell lung cancer (NSCLC)
Date and Time: April 11, 2022, 9:00 AM – 12:30 PM CT
Abstract: CT112
Developed in partnership with Genentech, a member of the Roche Group
Title: Application of interpretable neural graph network to predict gene expression signatures associated with tertiary lymphoid structures in histological images
Session Date and Time: April 11, 2022, 1:30 PM – 5:00 PM CT
Abstract: 1922
Posters will be available for registered attendees for on-demand viewing on the AACR website on April 8, 2022, at 1:00 PM ET.
“Our research demonstrates how PathAI’s algorithms can uncover novel insights or generate important molecular inferences from standard pathology samples that could not be identified using traditional manually-interpreted pathology or with more expensive and less accessible molecular approaches”, said Mike Montalto, Chief Scientific Officer at PathAI. “These translational insights are critically important to help accelerate drug development and get life-saving medicines to patients more quickly.”
A key example of this capability is highlighted in the poster, “AI-powered segmentation and analysis of nuclei morphology predicts genomic and clinical markers in multiple cancer types,” which describes PathAI’s machine learning-based nuclear segmentation model that associates features of nuclear morphology with clinically-relevant molecular and genomic markers across multiple tumor types.
The model was trained using over 29,000 annotations across multiple cancer types as well as five non-cancer tissues, to identify nuclei in any tissue type. Features related to nuclear shape, texture, and color were automatically extracted from the model and evaluated for their ability to predict genomic and molecular markers important for treatment selection and disease prognosis. When applied to breast, lung, and prostate tumor tissue, these models identified key combinations of nuclear features that were predictive of biomarkers including whole genome doubling, homologous recombination deficiency, HER2 positivity or elevated Gleason grade. These models could provide a way to rapidly genotype patients for selection of the most appropriate treatments without the need for extra tissue samples or sequencing protocols.
The full list of PathAI’s poster presentations is highlighted below. More information on each abstract can be found here.
Title: Quantification of TGFβ protein levels and digital pathology-based immune phenotyping reveal biomarkers for TGF-β blockade therapy patient selection in NSCLC
Session Date and Time: Online Only. E-poster available April 8, 2022, 12:00 PM CT.
Abstract: 5099
Developed in partnership with Sanofi
Title: Machine learning models identify histological features that can predict KEAP1 mutations in lung adenocarcinoma
Session Date and Time: April 10, 2022, 1:30 PM – 5:00 PM CT
Abstract: 449
Title: AI-powered segmentation and analysis of nuclei morphology predicts genomic and clinical markers in multiple cancer types
Session Date and Time: April 10, 2022, 1:30 PM – 5:00 PM CT
Abstract: 464
Title: AIM PD-L1-NSCLC: Artificial intelligence-powered PD-L1 quantification for accurate prediction of tumor proportion score in diverse, multi-stain clinical tissue samples
Session Date and Time: April 10, 2022, 1:30 PM – 5:00 PM CT
Abstract: 471
Title: AI-powered and manual assessment of PD-L1 are comparable in predicting response to neoadjuvant atezolizumab in patients (pts) with resectable non-squamous, non-small cell lung cancer (NSCLC)
Date and Time: April 11, 2022, 9:00 AM – 12:30 PM CT
Abstract: CT112
Developed in partnership with Genentech, a member of the Roche Group
Title: Application of interpretable neural graph network to predict gene expression signatures associated with tertiary lymphoid structures in histological images
Session Date and Time: April 11, 2022, 1:30 PM – 5:00 PM CT
Abstract: 1922
Posters will be available for registered attendees for on-demand viewing on the AACR website on April 8, 2022, at 1:00 PM ET.
Get In Touch
For media inquiries, please contact [email protected].