Generalizable system necessary for CDS systems’ success

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
 - Generalizable system necessary for CDS systems’ success

A generalizable system that can be applied to multiple cancer endpoints and validated using large multibatch datasets is necessary for the success of clinical decision support systems (CDSS), reported a review published in the October issue of the Journal of the American Medical Informatics Association.

Aimed at informing pathologists and informaticians with limited understanding of the key aspects of whole-slide imaging (WSI) analysis, the review was written by Sonal Kothari of the Georgia Institute of Technology, and colleagues.

Histopathological analysis is a standard clinical procedure that diagnoses the presence, type and progression of diseases like cancer. When diagnosing cancer patients using biopsy-derived tissue slides, pathologists must manually identify the most progressed regions. However, this can be incredibly time-consuming and subjective.

“CDSS and informatics methods can aid in decision-making by objectively quantifying morphological properties in histopathological images,” wrote Kothari and colleagues. “However, many of these systems and informatics methods still focus on images that represent only limited, manually selected regions of tissue slides rather than on whole-slide images.”

Thus, the review assesses histopathological WSI informatics, its associated challenges and the future research opportunities that exist to make CDSS a reality. Kothari and colleagues first examine current methods of histopathological WSI, such as quality control of histopathological images, feature extraction that captures image properties at pixel, object and semantic levels, predictive modeling that uses image features for diagnostic applications, and data visualization that explores WSI for de novo discovery.

After taking into account the current informatics methods that exist, challenges that have hindered the development of CDSS for WSI are discussed. These include: quality control; robust and fast image segmentation; knowledge (semantic-level) models for WSI; and ROI selection. The review then explores future research opportunities, such as the Cancer Genome Atlas. Lastly, it showcases a case study that features an example of a system that could make CDSS feasible.

“Currently, state-of-the-art software tools only provide limited image processing capabilities instead of complete data analysis for clinical decision-making. We aim to inspire researchers to conduct more research in pathology imaging informatics so that clinical decision support can become a reality,” wrote Kothari and colleagues.