Related Experiment Videos
Computer-derived nuclear features compared with axillary lymph node status for breast carcinoma prognosis
W H Wolberg1, W N Street, O L Mangasarian
1Department of Surgery, University of Wisconsin, Madison 53792, USA.
Cancer
|June 25, 1997
Summary
Computer analysis of nuclear features in fine-needle aspiration (FNA) samples offers a more accurate breast cancer prognosis than axillary lymph node examination, potentially reducing the need for dissection.
Area of Science:
- Oncology
- Pathology
- Medical Imaging Analysis
Background:
- Axillary lymph node status and tumor grade are key breast carcinoma prognostic factors.
- This study compared computer-derived nuclear features against lymph node involvement for prognostic accuracy.
Purpose of the Study:
- To evaluate the prognostic significance of computer-analyzed nuclear morphometric features from fine-needle aspiration (FNA) samples.
- To compare the accuracy of nuclear feature-based prognostication with traditional axillary lymph node assessment.
Main Methods:
- Analyzed 198 preoperative FNA samples from invasive breast carcinoma patients.
- Utilized a multivariate prediction model for distant recurrence based on nuclear features.
- Cross-validated prognostic predictions and compared accuracy with lymph node involvement.
Main Results:
- Nuclear feature stratification better distinguished the best and intermediate prognosis groups.
- Lymph node stratification was superior for separating intermediate and worst prognosis groups.
- Adding lymph node status or tumor size did not improve prognostic accuracy of nuclear features.
Conclusions:
- Computer analysis of preoperative FNA provides more accurate prognostication of favorable breast cancer patients.
- This method may eliminate the necessity for routine axillary lymph node dissection.