Association of stellate mammographic pattern with survival in small invasive breast tumors.

OBJECTIVE

We tested whether the stellate mammographic pattern of presentation of small breast tumors is a better indicator of survival than other patterns.

MATERIALS AND METHODS

Patients with primary invasive breast cancer diagnosed in 1993-1997 were grouped according to the size of the lesion: 0.1-0.9 cm and 1.0-1.4 cm. Each tumor was placed in one of five mammographic prognostic categories: stellate without calcifications; circular without calcifications; and calcifications with or without tumor mass in a casting, crushed-stone, or other (powdery, punctuate, or round) pattern. To assess reproducibility, a second radiologist gave an independent interpretation in the first 109 cases. Descriptive data were stratified by tumor size, and tests of association were done with an extension of Fisher's exact test. Odds ratios and confidence intervals were computed. Weighted log-rank test and Kaplan-Meier survival curves were used to compare breast cancer survival in the stellate group compared with the other groups.

RESULTS

Two hundred one consecutive patients with a median follow-up period of 7.4 years were identified. There were nine breast cancer deaths. The stellate morphologic pattern was most common (91 [45.3%] of 201 cases), yet there was only one breast cancer death in this group (survival rate, 98.9%; 95% confidence interval [CI], 96.7-100%). In the group of stellate lesions smaller than 1.0 cm, 67.6% (25/37) of the tumors were well-differentiated without lymph node metastasis (30 [96.8%] of 31 cases), and there were no deaths. In the group of stellate lesions measuring 1.0-1.4 cm, 66.7% (36/54) of the tumors were well-differentiated with a 19.6% risk of lymph node metastasis, one death, and a survival rate of 98.1% (52/53; 95% CI, 94.4-100%). Circular tumors accounted for 29.9% (60/201) of tumors and 55.6% (5/9) of breast cancer deaths. Casting and crushed-stone microcalcifications were associated with 33.3% (3/9) of disease-specific deaths. A weighted kappa value of 0.89 (CI, 0.83-0.94) indicated very high agreement of pattern assignment.

CONCLUSION

Stellate tumors had a significantly better survival prognosis than tumors with other patterns even though there were no differences in treatment. By recognizing these small malignant lesions, trained radiologists may be able to identify tumors that pose negligible risk of breast cancer death.