Abstract Book

S17

ESTRO 37

PV-0042 SBRT for peripheral lung tumors >5 cm: first results of the multicenter phase I/II VOLUMES trial H. Peulen 1,2 , J.J. Sonke 2 , E. Van Werkhoven 3 , M. Werner- Wasik 4 , I. Grills 5 , M. Giuliani 6 , A. Hope 6 , F. Mantel 7 , M. Guckenberger 7,8 , J. Belderbos 2 1 Catharina Hospital, Radiation Oncology, Eindhoven, The Netherlands 2 Netherlands Cancer Institute, Radiation Oncology, Amsterdam, The Netherlands 3 Netherlands Cancer Institute, Radiotherapy, Amsterdam, The Netherlands 4 Thomas Jefferson University Hospital, Radiation Oncology, Philadelphia, USA 5 Beaumont Hospital, Radiation Oncology, Royal Oak, USA 6 University of Toronto and Princess Margaret Cancer Center, Radiation Oncology, Toronto, Canada 7 University of Wuerzburg, Radiation Oncology, Wuerzburg, Germany 8 University Hospital Zurich, Radiation Oncology, Zurich, Switzerland Purpose or Objective To report the first results of the prospective multicenter phase I/II VOLUMES trial (NCT01543672) investigating the safety of SBRT for peripheral lung tumors >5 cm. Material and Methods Eligible patients had a peripheral lesion >5 cm (primary tumor or solitary metastasis) and were either inoperable or refused surgery. Patients with 2 tumors in close proximity (50% isodose line coverage of both PTVs, sum diameter >5 cm) were accepted. Radiation pneumonitis (RP) was anticipated to be the dose-limiting toxicity (DLT), which is correlated with the Mean-Lung Dose (MLD) (1). Therefore, an MLD (EQD 2 α/β=3 Gy) escalation was performed. Six risk groups (RG) were defined, increasing the MLD with 3 Gy in each step, ranging from <9 Gy (RG1) to 21-24 Gy (RG6) for the lowest and highest RG respectively. SBRT was delivered with VMAT and on-line image guidance. DLT was monitored by the Time-to-Event Continuous Reassessment Method (TITE-CRM) and defined according to the CTCAE v4. For non-lung toxicity a 3+3 design was used. The primary endpoint was to assess the maximum tolerated MLD with a 20% probability of DLT. Secondary endpoints were locoregional control, overall survival and quality of life. Results Between November 2011 and April 2016 all 30 patients were enrolled, with patient numbers per each RG as follows: RG1:7; RG2:8; RG3:10; RG4:3; RG5:2; RG6:0. Median age was 72 years (range 54-87), 67% was male. Most patients had a WHO score-1 with a median Charlson co-morbidity index of 2 (range 0-7). Primary NSCLC was diagnosed in 27 patients, of which 19 cT2b and 8 cT3 (7 th TNM). Three patients had lung metastases. Tumors were biopsy proven in 63%. The median GTV and PTV of large tumors (n=22) were 65cc (18-127) and 157cc (49-340) vs. 23cc (1-45) and 78cc (16-132) for 2 tumors (n=8). The most frequent (47%) SBRT schedule was 3x18 Gy. After a median follow-up (FU) of 19 months 3 DLTs were reported: one RP G5 at 8 months (MLD=19.5 Gy), one cavitating lung infection G3 at 6 months (MLD=18 Gy) and one G4 skin ulceration at 22 months. G3 toxicity was scored 5 times in 3 patients: dysphagia, fracture, chest wall pain, anorexia and dyspnea. The predicted maximum tolerated MLD was 21-24 Gy (95% probability interval 0.036-0.304) (Fig 1). One year local control (LC) (Fig 2), regional control (RC), distant control (DC) and disease free survival (DFS) were 96% (95%CI 0.9-1.0), 92% (95%CI 0.8-1.0), 78% (95%CI 0.6-1.0) and 76% (95%CI 0.6-1.0) respectively. Two year LC, RC, DC and DFS were 83% (95%CI 0.7-1.0), 84% (95%CI 0.7-1.0), 68% (95%CI 0.5-1.0) and 49% (95%CI 0.3-0.8) respectively. Overall survival (OS) at 1 and 2 years was 83% (95%CI 0.7-1.0) and 55% (95%CI 0.4-0.9) respectively. Fig 1

radiotherapy (SBRT) have an excellent oncological prognosis, but a high mortality rate because of co- morbidities. Therapeutic decisions are influenced by prognostic factors such as WHO Performance Status (PS), but refinement with more objective factors is warranted. Muscular strenght has been associated with mortality and morbidity in other diseases, and is easy to measure with hand grip strength (HGS). We therefore analyzed the value of HGS as a prognostic factor for survival in inoperable stage I NSCLC patients.

Material and Methods A prospective cohort study was done in medically inoperable (as defined at the multidisciplinary board) stage I patients. All patients had a HGS test (Jamar hydraulic hand dynamometer) before SBRT. They were instructed to grip the handle with maximal strength during 3 seconds, and this was measured 3 times in each hand. Handgrip weakness was defined as maximal handgrip strength below percentile ten of UK Biobank reference values, taking gender, age and height into account. To analyse differences in overall survival (OS) Cox proportional hazard model was used; differences are expressed in Hazard Ratio (HR) with 95 % confidence intervals. Results 226 stage I patients were included. Mean age was 72 ± 9 years. Mean Body Mass Index (BMI) was 25.0 ± 5.0 kg/m 2 . 59 % were males and 41 % females. In 45% of patients no pathology was obtained. WHO PS 0, 1, 2 and 3 was 16%, 60%, 22% and 3%, respectively. Handgrip weakness was seen in 31 % of the patients, whereas 69% had no weakness. At 5 years, univariate analysis showed a poorer survival for handgrip weakness (HR 1.78 (1.24-2.55), p=0.002) (Fig. 1) and WHO PS 2 or more (HR 1.49 (1.01-2.21), p=0.04). At multivariate analysis handgrip weakness (HR 1.52 (1.03-2.26), p=0.04), male gender (HR 1.80 (1.21- 2.66), p=0.004), higher age (per year or older) (HR 1.03 (1.00-1.05), p=0.02) and lower BMI (per kg/m 2 higher) (HR 0.93 (0.89-0.97), p=0.01) were associated with poorer survival at 5 years, whereas WHO PS 2 or more (HR 1.38 (0.92-2.09), p=0.12) was not significant anymore. Even at 1-year, handgrip was an independent prognostic factor for OS. Conclusion Handgrip strength is an objective, cheap and easy to measure independent prognostic parameter for short-and long-term OS in stage I NSCLC treated with SBRT.