Blog

Tumor-Type Prediction Using Targeted Clinical Genomic Sequencing Data

Photo of Madison Darmofal from mskcc.org

Jul 22, 2024, 08:58

Tumor-Type Prediction Using Targeted Clinical Genomic Sequencing Data

Maria Natalia Gandur Quiroga shared on X:

“Deep-Learning Model for Tumor-Type Prediction Using Targeted Clinical Genomic Sequencing Data

Results and Key Messages from the Article:

1. Deep-Learning Model GDD-ENS:

Developed to classify tumor types using genomic sequencing data.
Utilizes a comprehensive dataset of 39,787 solid tumors.

2.High Accuracy:

GDD-ENS achieves 93% accuracy for high-confidence predictions across 38 cancer types.
Comparable to whole-genome sequencing (WGS)-based methods.

3. Clinical Implementation:

GDD-ENS can be integrated into clinical workflows to guide treatment decisions in real-time.
Provides tumor type predictions and patient-specific feature importance values.

4. Model Features:

Hyperparameter ensemble classifier of deep neural networks.
Allows incorporation of patient-specific clinical information for enhanced predictions.

5. Impact on Challenging Cases:

Assists in diagnosing rare cancer types and cancers of unknown primary (CUP).
Improves diagnostic accuracy in challenging scenarios.

6. Prostate Cancer:

GDD-ENS includes prostate cancer in its predictions with high accuracy.
Provides valuable diagnostic information that can guide more effective treatments for prostate cancer patients.

Implications for Clinical Practice: The GDD-ENS model has the potential to transform clinical practice by providing more accurate and faster diagnoses, especially in cases where traditional methods are inconclusive. This can lead to more personalized and effective treatments, improving patient outcomes.

Tumor-Type Prediction Using Targeted Clinical Genomic Sequencing Data

Authors: Madison Darmofal, Shalabh Suman, Gurnit Atwal, Michael Toomey, Jie-Fu Chen, Jason C. Chang, Efsevia Vakiani, Anna M. Varghese, Anoop Balakrishnan Rema, Aijazuddin Syed, Nikolaus Schultz, Michael F. Berger, Quaid Morris.”

Tumor-Type Prediction Using Targeted Clinical Genomic Sequencing Data

Source: Maria Natalia Gandur Quiroga/X

Maria Natalia Gandur Quiroga is a Medical Oncologist and Chief of the Division of Genitourinary Medical Oncology at the Instituto de Oncología Ángel H. Roffo in Buenos Aires, Argentina. She is a Professor of Medicine at the University of Buenos Aires at the Oncologists Post Graduates Studies.

Her research focuses on clinical trials with aims to improve the treatment of patients with urologic tumors. She is an active member of the European Association for Cancer Research, Argentinian Medical Association and American Society of Clinical Oncology.

Aijazuddin Syed Anna M. Varghese Anoop Balakrishnan Rema cancer Efsevia Vakiani Gurnit Atwal Jason C. Chang Jie-Fu Chen Madison Darmofal María Natalia Gandur Quiroga Michael F. Berger Michael Toomey Nikolaus Schultz OncoDaily Oncology prostate cancer Quaid Morris Shalabh Suman

Mark Ruffalo and Brain Cancer: How He Went Against, How He Survived, and More

Sharon Osbourne and Colon Cancer: How She Went Against, How She Survived, and More

Marcia Cross and Anal Cancer: How She Went Against, How She Survived, and More

Pancreatic Cancer: Symptoms, Causes, Types, Diagnosis & Treatment

FDA approves Methotrexate for Pediatric acute Lymphoblastic Leukemia

Special Coverage