Padula, Gilbert D. A.

104 Background: Lymphedema (LE), a side effect of BC treatment, affects 8%–56% of women. The study tested the effectiveness of an education only vs education + exercise intervention on LE incidence in a group randomized trial (NCT00376597) of 38 cooperative group sites. Methods: Newly diagnosed women with stage I-III BC were consented and arm circumference and range of motion were measured pre-surgery. After surgery, patients (pts) were enrolled based on the randomization assignment of their treating institution, stratified by their type of node dissection (sentinel or axillary), and received either: education only: its causes, signs, symptoms and risk reduction strategies, or education + personalized exercise instruction from a physical therapist, with elastic compression garments to wear during heavy arm use and air travel. Pts completed surveys and arm measurements at baseline (BL), 12 and 18 months with the LE endpoint defined as ≥ 10% difference in limb volume at any time from BL to 18 months or physician diagnosis of LE. Cochran-Mantel-Haenszel (CMH) tests were used to compare the LE-free rate between arms, adjusted for type of biopsy. Logistic regression assessed predictors of LE. Results: 568 pts were enrolled, 45% in education only, and 55% in education + exercise, with 554 (98%) included in the analyses. Pts were on average 57.7 years, White (83%), non-Hispanic (92%) and married (62%). 47% only had sentinel node dissection, 35% had mastectomy, and 37% had chemotherapy. The LE-free rate was 58% in education only vs 55% in education + exercise (CMH p = 0.73). Predictors of LE included surgery type (lumpectomy vs mastectomy, OR = 2.22, p < 0.001) and chemotherapy (no vs yes, OR = 1.73, p = 0.02), after adjusting for BMI and type of node dissection. Only one-third wore the sleeve at least 75% of the time and only 50% did LE exercises at least weekly. Conclusions: There was no difference by intervention arm in LE incidence by 18 months. Poor adherence in the education + exercise arm may have been a factor. Additional analyses will assess quality of life, body image and knowledge about LE by arm. Further research is needed to effectively prevent LE in pts after BC surgery. Research Support: UL1TR001409; U10CA037447, UG1CA189823 Clinical trial information: NCT00376597.

347 Background: 9,704 was the first phase 3 pancreatic cancer trial to validate the prognostic value of postresection CA19-9 for overall survival (OS), with values > 90/180 associated with worse OS. All pts received adjuvant gemcitabine or 5-FU and chemo-radiotherapy (RT). This analysis evaluates patterns of disease failure. Methods: SMS was negative, positive, or unknown. CA19-9 was analyzed at cut points 90, 180 and continuously. LRF and DF were estimated by cumulative incidence and Gray’s test compared. Cox hazard models were used for multivariate analyses (MVA) and included treatment, tumor site, size and nodal status. To adjust for multiple comparisons a p-value < 0.01 is statistically significant and 0.01 to < 0.05 a trend. Results: 538 pts accrued, with 451 eligible and analyzable for SMS and 385 for CA19-9. For CA19-9, 132 (34%) were Lewis Antigen negative (no CA19-9 expression), 200 (52%) < 90 and 220 (57%) < 180. 188 (42%) had negative margins, 152 (34%) positive and 111 (25%) unknown (i.e., no margin comment in path report; shown to have outcomes similar to negative margin pts). Pts with CA19-9 ≥ 180 were more likely to have tumors ≥ 3 cm and pts with positive SMS more likely to have KPS 60 - 80, T3/T4, or N1 disease. On univariate analysis (UVA) CA19-9 cut at 90 was associated with significant increases in both LRF (trend) and DF; in the gemcitabine arm this was seen in DF, not in LRF; in the 5-FU arm it was seen in both. Results were similar at the 180 cut point and continuously. SMS on UVA was not associated with increase in LRF/DF; see Table. On MVA, CA19-9 > 90 was significantly associated with LRF and DF; positive SMS showed only a trend for DF. Conclusions: Postresection CA19-9 has significant association with both LRF and DF not seen with SMS. These findings support continued use of RT in trials and consideration of dose intensification among pts with elevated postresection CA19-9. Grants: NCI U10, CA21661, CA37422, CA180868, CA180822. Clinical trial information: 0000000. [Table: see text]