Go, Ronald S.

9044 Background: Prior studies have shown superior surgical outcomes of stage I-III non-small cell lung cancer (NSCLC) in centers with higher patient volumes. However, there is a lack of such information in stage IV NSCLC. In this study, we aim to determine the association between the number of patients with stage IV NSCLC treated annually at a treatment facility (volume) and all-cause mortality (outcome). Methods: Using the National Cancer Database, we identified patients diagnosed with stage IV NSCLC between 2004 and 2013. We classified the facilities by quartiles (Q; mean patients with NSCLC treated per year): Q1: < 13.8; Q2: 13.8 to 23.6, Q3: 23.6 to 30.3, and Q4: > 30.3. We used sandwich variance estimators to account for clustering of patients within facilities and Cox regression to determine the volume-outcome relationship, adjusting for demographic (sex, age, race), socioeconomic (insurance type), receipt of chemotherapy, and comorbid (Charlson-Deyo score) factors and year of diagnosis. Results: There were 281,654 patients with stage IV NSCLC treated at 1,275 facilities. The median age at diagnosis was 66 years, and 55.7% were men. The median annual facility volume was 23.6 patients per year (range, 1.0 to 301.4). The distribution of patients according to facility volume was: Q1: 6.6%, Q2: 14.9%, Q3: 25.4%, and Q4: 53.1%. The unadjusted median overall survival by facility volume was: Q1: 4.4 months, Q2: 4.5 months, Q3: 4.7 months, and Q4: 5.3 months ( P< .001). Multivariable analysis showed that facility volume was independently associated with all-cause mortality. Compared with patients treated at Q4 facilities, patients treated at lower-quartile facilities had a small but significantly higher risk of death (Q3 hazard ratio [HR], 1.05 [95% CI, 1.03 to 1.07]; Q2 HR, 1.06 [95% CI, 1.03 to 1.09]; Q1 HR, 1.09 [95% CI, 1.06 to 1.13]). Conclusions: Patients who were treated for stage IV NSCLC at lower-volume facilities had a significantly higher risk of all-cause mortality compared with those who were treated at lower-volume facilities. [Table: see text]

e19532 Background: The diagnosis of amyloidosis requires histological confirmation of Congo-red (CR) deposits. The tissue source is preferably fat and/or bone marrow biopsy, but at times organ biopsy is required. Methods: We studied 612 patients with systemic light chain amyloidosis to characterize the tissues used to establish the diagnosis Results: The median number of tissue samples biopsied for CR staining was 3. Ninety-five percent of bone marrow (BM) biopsies were stained for CR, while 79% of patients had fat aspiration performed for CR staining. Overall, 76% of patients underwent both procedures. CR stain sensitivity was 69% in BM samples, 75% in fat aspiration and in 89% for both sources combined. CR stain sensitivity was 97-100% for heart, renal and liver biopsies. 42% of patients with renal involvement, 21% of patients with liver involvement and 13% of patients with heart involvement underwent organ biopsy, when a less invasive safer biopsy would have established the diagnosis. Predictors of need for organ biopsy were male sex, limited organ involvement and lack of fat aspiration (Table). Conclusions: Fat aspiration is underutilized for histologic confirmation of amyloidosis. Organ biopsies (particularly renal biopsies) were performed at a high rate, which represents a failure to recognize the disease (i.e. failure to screen with a light chain assay etc.). Early awareness of amyloidosis in the differential diagnosis of patients with organ dysfunction may lead to fewer unnecessary organ biopsies [Table: see text]

e19530 Background: Neuropathy is a universal symptom of patients with POEMS syndrome, which makes the use of bortezomib less appealing. However, proteasome inhibitors (PIs) can provide rapid and deep hematologic responses in plasma cell disorders. Because there is very little data regarding the use PIs in POEMS syndrome, we performed a study of patients with POEMS syndrome treated with PIs. Methods: Among the 167 patients with POEMS syndrome seen at our institution between 2001-2016, we identified 12 patients who received a PI. One patient receiving < 1 cycle of bortezomib because of worsening neuropathy was included but was excluded from response analyses. One patient received 2 different PIs at separate points during the disease course and each course was considered separately. Four response categories were considered: hematologic, PET, VEGF and clinical as previously described (Dispenzieri et. al AJH 2015). Results: Among the 12 patients who received 13 courses of therapy, the median age at PI initiation was 59; 8 patients were male. Median number of prior therapies was 1 (0-4) and median time from diagnosis to initiation of PI was 75 months. Excluding steroids, regimens were as follows: bortezomib, n=3; bortezomib and cyclophosphamide, n=4; carfilzomib, n=1; and ixazomib, n=1. With a median follow-up from PI initiation of 18 months, the median duration of treatment was 7 months (1-18). Three patients achieved a hematologic CR. Of 6 patients with an abnormal baseline VEGF, 3 achieved normalization VEGF levels. Of 5 patients with an abnormal baseline PET, 2 achieved normalization of their PET. Two patients had improvement of their neuropathy, with the rest having stable symptoms. Of 9 patients with fluid overload, 8 improved. Only 2 patients stopped therapy due to worsening neuropathy related to PIs. Median PFS and OS from initiation of PI were 50 months and 53 months, respectively. Conclusions: We show that PI use in patients with POEMS syndrome can be safe and effective. Nearly 75% of patients treated with bortezomib had no exacerbation of their neuropathy. Clinical improvement in severe fluid overload symptoms (edema, ascites, pleural effusions) is almost universal, making PI-based therapy an attractive potential option in these patients.

8049 Background: Complete response in AL is defined as normal FLC ratio with negative serum and urine immunofixation. It is not clear if high involved serum FLC (hIFLC) in a patient with normal ratio may contribute to ongoing amyloid formation and hence affect outcomes. Methods: Data of 1308 patients (pts) with systemic AL seen within 90 days of diagnosis, at Mayo Clinic between 2006-2015, was analyzed retrospectively. Among these, 369 pts had 2 consecutive normal FLC ratio values after 1st line treatment and form the study population. Log rank test was used to estimate survival differences. Results: Among these 369 pts, pts with hIFLC at 1streading of normal FLC ratio (hIFLC1; n=170; 46.1%) were compared to those who did not (n=199; 53.9%). At diagnosis, the median age [61.5 vs 60.8 years (y); p=0.2], proportion of males (62.4 vs 58.3%; p=0.4), percentage of pts with renal involvement (73.5 vs 64.8%; p=0.07), in mayo stage I / II / III / IV (32.9% / 23%/ 27.3% / 16.8% vs 43.6 %/ 22.9% / 18.1% / 15.4%; p=0.1), with bone marrow plasma cells >10% (24.2 vs 30%; p=0.2) and with presence of t(11; 14)(48.4 vs 60; p=0.08) was similar, while cardiac (67.5 vs 53.3%; p=0.006) and hepatic (18.2 vs 9.1%; p=0.01) involvement was higher in hIFLC1 group. The median follow-up from diagnosis was 6.1 y (95% CI; 5.6, 6.8). The median progression free survival (PFS) in pts who had hIFLC1 was lower than for those who did not; 2.6 y (95% CI; 1.9, 4.5) vs 5.2 y (95% CI; 4.6, 6.4), p<0.0001, as was the median overall survival [OS; 6.7 y (95% CI; 4.5, 8.3) vs not reached (NR), p<0.0001]. We performed a more stringent comparison for pts with 2 consecutive hIFLC values (hIFLC2; n=112; 30.4%) versus not (n=257; 69.6%). The median PFS (3.2 y; 95%CI; 2.2, 4.5 vs 5.6 y; 95% CI; 4.7, 7.1; p<0.0001) and OS (7.8 y; 95% CI; 6.4, NR vs NR; 95%CI; 9.5, NR; p<0.0001) were significantly reduced in pts with hIFLC2 versus not as well. A multivariate analysis confirmed an impact of hIFLC1 and hIFLC2 on PFS/OS independent of serum creatinine. Conclusions: In pts with systemic AL, persistent elevation of the involved FLC predicts for poor prognosis (independent of serum creatinine) even among those who achieved normal FLC ratio after 1st line treatment.

8048 Background: Organ response (OR) in AL is often delayed and difficult to predict early. Methods: We retrospectively analyzed 1308 patients (pts) with newly diagnosed AL from 2006 – 2015 to determine factors which could predict for OR. Results: Median age was 64 years (yr) and Mayo Stage was: 1 (22%); 2 (23%); 3 (25%); 4 (31%). Organ involvement was: cardiac (74%, n=932); renal (59%, n=738), liver (16%, n=205); gut (24%, n=310) and autonomic (12%, n=152). 59% (n=765) had > 1 organ involved, including 43% (n=567) with > 1 critical organ (heart, kidney, liver) involved. Treatment was: ASCT based (28%, n=330, N=1186), bortezomib based (24%, n=281), alkylator based (33%, n=392), others (5%, n=54) and none (10%). In evaluable pts, VGPR or better rates were: 53% at 6 months (m) (N=625), 72% at 12 m (N=465) and 57% overall (N=688). Table 1 lists OR at various time points. Complete OR in all involved critical organs was seen in: 51% (n=308, N=600), partial response (at least 1 OR when >1 organ involved) in 12% (n=73) and none in 37% (n=219). Complete OR was associated with better overall survival (OS) than partial or no OR (median OS: not reached vs 42 m vs 29 m; P <0.0001). In multivariate model the following variables at baseline or 1 yr mark were predictive of complete OR: lower Mayo Stage (p=0.01), fewer critical organs involved (p=0.007), higher baseline GFR (p=0.03), female sex (Complete OR 60% vs 47%; p=0.04) and VGPR at 1 yr (Complete OR 70% vs. 36%; p <0.0001). Other factors included in the model were age (p=0.9), bilirubin (p=0.1) and transplant (p=0.2). All aforementioned factors were significant in univariate analysis. Conclusions: Achievement of response in all involved critical organs is associated with better survival in AL pts than partial or no OR. Various baseline factors and VGPR at 1 yr can predict for achieving complete OR, with 70% pts who achieve VGPR at 1 yr having a complete OR. [Table: see text]

8047 Background: Presence of CPCs by six color MFC is associated with worse outcomes in multiple myeloma (MM). Using a slide based immunofluorescence assay, CPCs are associated with worse prognosis in AL, but outcomes with MFC are not known. Methods: We retrospectively analyzed 154 patients (pts) from Jan 2008 – Dec 2015 with AL who had CPCs analyzed by MFC at diagnosis. Results: CPCs were present in 42% pts (n = 64, median = 88 per 150,000 events, range 6-17844). Table 1 lists baseline characteristics, treatment and response. In univariate analysis, bone marrow (BM) plasma cells > 15% (p<0.0001), dFLC > 60 mg/dL (p=0.02) and presence of active MM (p=0.0004) were associated with detectable CPCs. In multivariate (MV) model, only BM plasma cells > 15% predicted for presence of CPCs (p=0.02). There was no difference in hematologic or organ response in the 2 groups. Median follow up was 43 months (m). Pts with detectable CPCs had worse survival than those without detectable CPCs (90 m vs. 98 m, p=0.0004). In MV analysis model with Mayo Stage and detectable CPCs, presence of CPCs at diagnosis was independently associated with poor survival (CPCs p= 0.02, Mayo Stage p < 0.0001). Conclusions: Presence of detectable CPCs is an independent adverse prognostic factor for survival in AL. Comparatively fewer pts with detectable CPCs underwent early ASCT, which may reflect a more aggressive disease presentation. [Table: see text]

8038 Background: Daratumumab-based Combination Therapies (DCT) with bortezomib (V)/ lenalidomide (R)/ pomalidomide (P) and dexamethasone (d) showed exceptional activity in RRMM in trials. Experience outside of trials since the approval of Daratumumab (D) in 2015 is limited. Methods: RRMM pts seen at Mayo Clinic, MN from 12/2015 -12/2016 were reviewed. Pts who received ≥ 1cycle of DCT were included. Time-to-event analyses were done from date of starting DCT. Common terminology criteria for adverse events v4.0 were used to grade toxicities. Results: Of 130 pts, 59% were males and median age at DCT initiation was 67 (43-93) years, ECOG performance score was ≥2 in 29%. Pts were classified as mSMART high (22%), intermediate (22%) or standard (56%) risk. Median time from diagnosis to initiation of DCT was 51.3 (5-156) months (m), and median number of prior therapies was 4 (1-14). 14% of pts were refractory to prior D monotherapy. Fifty-three (41%), 34 (26%) and 25 (19%) received DPd, DRd and DVd respectively. Eighteen (14%) pts received ‘other’ DCT. Median time to first response (≥ PR) was 3.1 m (95% CI 2.1-4.6). Overall response rate was 46%, [CR-2%, VGPR-18%, PR-26%]. Minimal response was seen in 17%, with clinical benefit rate of 62%. Median estimated follow up from initiation of DCT was 5.5 m (CI 4.2-6.1). The median duration of response was 6.1 m [CI 5.1- not reached (NR)]. Median progression free survival (PFS) was 5.5 m (CI 4.1-7.8) and median time to next therapy (TTNT) was 5.9 m (CI 4.6-9.4). Median PFS for DPd, DRd, DVd and other DCTs were 4.6 (CI 2.7-NR), 7.8 (CI 5-NR), 3.9 (CI 2.1-NR) and 3.9 (CI 2.8-8.2) m, respectively (p = 0.3). Median PFS for quadruple refractory (n = 28) MM was 2.8 m (CI 2.2-5.3) vs 5.9 m (CI 4.9-NR) for the rest (p < 0.01). Median overall survival (OS) from DCT was NR (CI 11.4-NR). Grade 3 or higher hematological toxicities were seen in 42% of pts. Other toxicities included infections (37%), fatigue (31%), infusion reactions (16%) and diarrhea (10%). Conclusions: DCT are effective in RRMM, but the PFS remains short particularly in quadruple refractory pts, reflecting the challenges encountered in managing heavily-pretreated, and often less fit patients, in routine practice.

8032 Background: Human adipocytes can contribute directly to the in vitro growth and progression of multiple myeloma (MM) cell lines. Clinically, an elevated body mass index (BMI) has been associated with an increased risk of MGUS and a shorter time to progression (TTP) of MGUS to MM. However, the impact of BMI on the risk of early progression to MM from a more advanced plasma cell disorder such as smoldering MM (SMM) remains unknown. Methods: This study included patients (pts) with a known or new diagnosis of SMM evaluated at the Mayo Clinic, Rochester from January 2000-December 2010. Pts were classified based on their BMI as: normal (< 25) and elevated (>/= 25) BMI. Progression to symptomatic MM was defined by the development of hypercalcemia, renal insufficiency, anemia or lytic bone lesions. Results: There were 306 pts with a diagnosis of SMM who were included in this analysis. The median follow up was 106 months. There were 203 (66%) pts who progressed to symptomatic MM at last follow up. The median BMI of the group was 27.5 (Range: 17.2 – 56.4). There were 228 (75%) pts with an elevated BMI. There were 76 (28%) pts who had myeloma defining events (MDEs) such as a serum free light chain ratio > 100 or > 60% clonal bone marrow plasma cells at initial evaluation. MDEs were present in 17% and 33% of pts with a normal and elevated BMI respectively (P = 0.011). The median TTP of SMM to MM in pts with a normal and elevated BMI was 64 and 36 months respectively (P = 0.0006). The 2-year progression rate of SMM to symptomatic MM in pts with a normal and elevated BMI was 16% and 42% respectively (P < 0.0001). Upon limiting the analysis to only SMM pts without MDEs at initial evaluation (N =187), the 2-year progression rate to symptomatic MM with a normal and elevated BMI was 15% and 33% respectively (P = 0.013). In a multivariable model, only elevated BMI (P = 0.004) and increasing clonal bone marrow plasma cells (P = 0.001) was statistically significant in predicting for a 2-year progression to MM. Conclusions: SMM pts with an elevated BMI appear to have a higher risk of early progression to MM than those with a normal BMI. This study provides evidence of a potentially modifiable risk factor for the progression of SMM to MM and warrants confirmation in larger studies.