Inventors:
- Albany NY, US
- Austin TX, US
- Atlanta GA, US
- Seattle WA, US
Rebecca BATISTE - Stony Brook NY, US
Le HOU - Stony Brook NY, US
Vu NGUYEN - Stony Brook NY, US
Dimitrios SAMARAS - Rocky Point NY, US
Arvind RAO - Houston TX, US
John VAN ARNAM - Houston TX, US
Pankaj SINGH - Houston TX, US
Alexander LAZAR - Houston TX, US
Ashish SHARMA - Atlanta GA, US
Ilya SHMULEVICH - Seattle WA, US
Vesteinn THORSSON - Seattle WA, US
International Classification:
G06T 7/00
G06K 9/32
G06T 9/00
G06K 9/62
G06T 7/11
Abstract:
A system associated with quantifying a density level of tumor-infiltrating lymphocytes, based on prediction of reconstructed TIL information associated with tumoral tissue image data during pathology analysis of the tissue image data is disclosed. The system receives digitized diagnostic and stained whole-slide image data related to tissue of a particular type of tumoral data. Defined are regions of interest that represents a portion of, or a full image of the whole-slide image data. The image data is encoded into segmented data portions based on convolutional autoencoding of objects associated with the collection of image data. The density of tumor-infiltrating lymphocytes is determined of bounded segmented data portions for respective classification of the regions of interest. A classification label is assigned to the regions of interest. It is determined whether an assigned classification label is above a pre-determined threshold probability value of lymphocyte infiltrated. The threshold probability value is adjusted in order to re-assign the classification label to the regions of interest based on a varied sensitivity level of density of lymphocyte infiltrated. A trained classification model is generated based on the re-assigned classification labels to the regions of interest associated with segmented data portions using the adjusted threshold probability value. An unlabeled image data set is received to iteratively classify the segmented data portions based on a lymphocyte density level associated with portions of the unlabeled image data set, using the trained classification model. Tumor-infiltrating lymphocyte representations are generated based on prediction of TIL information associated with classified segmented data portions. A refined TIL representation based on prediction of the TIL representations is generated using the adjusted threshold probability value associated with the classified segmented data portions. A corresponding method and computer-readable device are also disclosed.