IMAGING IN THE EVALUATION OF PRIMARY CHEMOTHERAPY IN EWING/PNET TUMORS. CORRELATION WITH NECROSIS AND SURVIVAL. AN ITALIAN SARCOMA GROUP/SCANDINAVIAN SARCOMA GROUP STUDY

P. Picci, I. Taksdal, MC Malaguti, C. Ferrari, G. Bacci, G. Saeter, C. Monti, T. Alvegard.

In Ewing/PNET tumors the grade of necrosis histologically evaluated after primary chemotherapy is now unanimously considered the most important prognostic factor, and it is justified to intensify chemotherapy in those patients with a poor histological response. Because not all the patients with this disease can be treated with surgery as local treatment, this important prognostic factor cannot be used in same cases. The present study was conducted to evaluate if the comparison between CT and/or MRI performed before and at the end of primary chemotherapy can supply information predictive of response and useful as prognostic factor.

55 patients treated in the years 1985-1996 with localized disease were available for review. 25% had tumors located in the axial skeleton. All had received neoadjuvant chemotherapy following protocols in use at the time, and all had surgery as local treatment. The evaluation was based on the variation of the soft tissue component at consecutive CT or MRI. Only the total disappearance or the total ossification/calcification of the soft tissue component was considered a good radiological response. Histologically, as reported previously (JCO, 1997), a total necrosis or the persistence of microscopic foci were considered a good response.

22 cases (40%) had a good radiological response and 28 (51%) had a good histological response. A good histological response was detected in 18 of the 22 cases with good radiological response (82%) in comparison to 10 of the 33 with poor radiological response (30%) (p=0.005). A total concordance radio/histological (good/good and poor/poor) was found in 18 and 23 patients respectively (33% + 42% = 75%). In 10 cases (18%) a poor radiological response corresponded to a good histological necrosis and viceversa in 4 cases (7%). Regarding prognosis, 21 of the 22 cases with good radiological response (95%) are continuosly disease free in comparison to 12 of the 33 with poor radiological response (36%) (p=0.0004). An additional 8 patients, not having a soft tissue invasion at diagnosis, could not be evaluated with the reported criteria. All these 8 patients are continuously disease-free, supporting the hypothesis that soft tissue invasion (and its persistence after induction chemotherapy) is a bad prognostic factor. In conclusion, the method seems to be reliable and simple enough to be used especially within multicentric studies. It can be employed to differentiate maintenance chemotherapy in those patients not receiving surgery as local treatment. This method will be used in the new forthcoming Italian Sarcoma Group/Scandinavian Sarcoma Group (ISG/SSG III) joint protocol.