Embedded Files
General Conclusions
General Conclusions
Our model achieved good diagnostic accuracy with a peak accuracy of 95%.
Our model achieved good diagnostic accuracy with a peak accuracy of 95%.
Increasing the size of the training data set led to improved accuracy.
Increasing the size of the training data set led to improved accuracy.
This is a promising method with potential to accurately detect differences between metabolic subtypes of TNBC which can eventually lead to better outcomes and future clinical applications.
This is a promising method with potential to accurately detect differences between metabolic subtypes of TNBC which can eventually lead to better outcomes and future clinical applications.
Future Directions
Future Directions
Collect more data to train the model with
Collect data using a more varied set of imaging collection parameters to improve the general applicability of the model
Train/test the model on data collected directly from patients.
Standards Used
Standards Used
This project, being conducted through Shi Lab at UCSD, followed IACUC guidelines and approval on ethical animal testing, as well as respective NIH/OLAW standards for vertebrate testing.
Standards that we followed during the data acquisition process fell under UCSD’s wet laboratory health and safety policies and Laboratory Safety Manual which ensured the safety of team members in the lab and reduced the risk of outside contamination for the experiment (not as applicable for the actual imaging portion as it is for sample preparation).
For specific guidelines required for calibrating and operating the spontaneous Raman microscope, we followed rules set by the American Society for Testing and Materials (ASTM E 1840) for Raman Spectroscopy (taught to us by our mentors) as well as those set by the National Institute of Standards and Technology (NIST).
Page developed by Dmitriy Major
Page updated
Report abuse