Scientific Premise:

To develop an artificial intelligence system capable of augmenting / speeding up real-time margin assessment for complete tumor resection.

Motivation:

Complete surgical excision of tumor by itself or in conjunction with or as a precursor to chemotherapy is often required for the treatment of many solid tumors. This procedure requires assessment of tumor margins (tumor’s “edge”) to decide whether tumor/margins have been cleared. Typical excisions perform margin assessment post-operatively and use breadloafing as a sectioning technique, which only samples a small fraction of the tissue (1-2%). This leads to positive (detecting tumor post-operatively at margin) or missed (false negative) margins, which may lead to a repeat procedure or tumor recurrence respectively. Real-time total margin analysis forms the basis of Mohs Micrographic Surgery (MMS) avoiding post-operative positive margins of permanent sections and false negative margins secondary to the sampling bias of breadloafed grossing. Challenges to instituting a MMS approach to treat additional tumors types include enlisting providers that are capable of performing surgery, reading histology, and mapping tumor all in a timely manner. This can be circumvented by creating a seamless virtual workspace acting as a non-autonomous decision aid to facilitate these tasks. Our team is working to develop such a system, using Basal Cell Carcinoma as a target system, which could potentially reimagine the role of real-time margin analysis on larger and more complex tumor types.

Manuscripts: Arxiv coming soon!