Search

The Van Nuys prognostic index (VNPI) was thought to be useful for predicting response to radiotherapy and local recurrence of ductal carcinoma in situ (DCIS). We applied the VNPI under the conditions defined by Silverstein et al., in 367 retrospective DCIS entirely sectioned into serial macroscopic 2 mm slices (155 patients had radiotherapy, median follow-up 71 months). The percentage of positive blocks with DCIS was also estimated for each specimen with cut-offs at 30% and 60% to obtain three scores. One hundred and ninety five lesions had a low VNPI, 152 an intermediate VNPI, and 20 a high VNPI. There were 9% of local recurrences (half invasive, all in the group without radiotherapy) in the low VNPI group. The local recurrence rate increased with size (p=0.001), with reduction of distance to margins (p=0.05), with histologic grade (p=0.02), with percentage of positive blocks (p=0.0003) and with VNPI score (p=0.03). The percentage of positive blocks was the only independent predictor for local recurrence (p=0.0001). Conclusion: (1) The VNPI was a local recurrence rate predictor between the low and the intermediate groups but in our series the low VNPI group had a surprisingly high local recurrence rate. (2) Only prospective studies will assess the importance of margin width and the role of radiotherapy in maintaining local control. (3) Estimation of the percentage of positive blocks is simple, may be an alternative when measurement of DCIS is difficult and should be taken into account.

From 1980 to 1984, we observed 144 undifferentiated cancers, of which 116 (0.9% of all cancers treated during this period) were treated in our center. Immunohistochemical study classified 130 tumors (90%), which comprised 82 non-Hodgkin's lymphomas (57%), 32 carcinomas (22%), 7 melanomas (5%), 7 sarcomas (5%) and 2 others (1%). Sixty-nine patients, with the diagnostic problem of non-Hodgkin's lymphoma versus carcinoma, which was solved by immunohistology, were followed up for 44 ± 20 months. Lymphomas (57 cases) had a better clinical course than carcinomas (11 cases), and a clinical course similar to high-grade lymphomas identified by conventional histology. For 66 patients with the same problem, the treatment was started before the immunohistochemical diagnosis. This treatment was inappropriate in the light of the correct diagnosis in 16 of 66 cases (24% of all cases).