Predicting Metastasis in Breast Cancer: Comparing a Decision Tree with Domain Experts
110 consecutively diagnosed breast-cancer patients in all stages were included in a study to evaluate a selective surgical and radiotherapeutical treatment. The surgical treatment was total mastectomy and exploration of the axilla, with lymphnode biopsy and peroperative cytological examination. Axillary dissection was done only when this examination showed metastases. No radiotherapy was given to the axilla in patients with lateral cancers in the absence of metastases, or with limited metastasization (no periglandular growth, no growth in apical nodes). In medial and central cancers, radiotherapy was applied to the parasternal and supraclavicular nodes irrespective of axillary involvement. A staging system with a combined clinical and histopathological classification was used and formed the basis for the selective treatment. The corrected 5-year survival for the whole material was 80%, for those without axillary metastasis (Stage I) 95% and for those with axillary metastasis (Stage II) 68%. Six women were alive with known distant metastases. Of 63 patients without identified axillary metastases at the time of surgery, axillary recurrences occurred in only 3 (5%). It was concluded that patients without axillary metastases can be reliably selected by the peroperative examination used, and that in this group simple mastectomy results in a high disease-free survival. Early diagnosis and a possible beneficial effect of the actual therapeutic programme might both have contributed to the high overall survival.