Virtual Library
Start Your Search
M. Lambrecht
Author of
-
+
P1.13 - Radiology/Staging/Screening (ID 699)
- Event: WCLC 2017
- Type: Poster Session with Presenters Present
- Track: Radiology/Staging/Screening
- Presentations: 1
- Moderators:
- Coordinates: 10/16/2017, 09:30 - 16:00, Exhibit Hall (Hall B + C)
-
+
P1.13-007 - Is Central Lung Tumor Location Really Predictive for Occult Mediastinal Nodal Disease in (Suspected) NSCLC Staged cN0 on PET-CT? (ID 8779)
09:30 - 09:30 | Author(s): M. Lambrecht
- Abstract
Background:
Based on a 20-30% prevalence of occult mediastinal disease, current guidelines recommend preoperative invasive mediastinal staging in patients with central tumour location and negative mediastinum on PET-CT. A uniform definition of central tumour location is lacking. Our objective was to determine the best definition in predicting occult mediastinal disease.
Method:
A single institution prospective database was queried for patients with (suspected) NSCLC staged cN0 after PET-CT and referred to invasive staging and/or primary surgery. We evaluated 5 definitions of central tumour location (table 1).
Result:
Between 2005 and 2015, 822 patients were eligible. Radio-occult lesions were excluded from analysis (n=9). Preoperative histology was NSCLC in 49% and unknown in 51%. The lesion was subsolid in 7%. Tumour stage was cT1, cT2, cT3 and cT4 in 43%, 28% 17% and 11%, respectively. Invasive mediastinal staging (EBUS and/or mediastinoscopy) was performed in 31%. Surgical resection was performed in 97%, a median of 5 (IQR 3-6) nodal stations were examined. The final pathology was squamous NSCLC, non-squamous NSCLC, or other in 38%, 54% and 7%, respectively. Any nodal upstaging was found in 21% (13% pN1 and 8% pN2-3). Central tumour location demonstrated, compared to peripheral location, a 4 times higher risk for any nodal upstaging but not for N2-3 upstaging (table 1).
Conclusion:
When modern PET-CT fusion imaging points at clinical N0 NSCLC, the prevalence of occult mediastinal nodal disease was only 8% in our patient cohort. None of the five definitions of centrality we studied was predictive for occult pN2-N3. Overall nodal upstaging was 21%, however, and all definitions of centrality then had discriminatory value. These data question whether the indication of preoperative invasive mediastinal staging should be based on centrality alone. Table 1 Figure 1
-
+
P3.14 - Radiotherapy (ID 730)
- Event: WCLC 2017
- Type: Poster Session with Presenters Present
- Track: Radiotherapy
- Presentations: 1
- Moderators:
- Coordinates: 10/18/2017, 09:30 - 16:00, Exhibit Hall (Hall B + C)
-
+
P3.14-011 - Mean Heart Dose Is an Independent Risk Factor for Early Mortality After Chemoradiotherapy Treatment for Lung Cancer (ID 10342)
09:30 - 09:30 | Author(s): M. Lambrecht
- Abstract
Background:
Early mortality after (chemo)radiotherapy can be caused by treatment-related toxicities and thus by delivered doses to normal lung and heart tissues. However, prediction models for mortality incorporating dosimetry are lacking. This study explores the prognostic value of common dosimetric features.
Method:
Two prospective cohorts containing 218 and 181 curatively treated stage I-III lung cancer patients from 2003-2007 and 2013-2016 periods, respectively, were studied. Prescribed dose was 66Gy/2Gy (concurrent chemotherapy), 66Gy/2.75Gy (sequential or no chemotherapy) or a similar schedule. Clinical (WHO performance status, age, T stage, N stage and primary gross tumor volume (GTV)) and dosimetric (mean lung dose (MLD) and mean heart dose (MHD)) covariates were analysed. Cox regression models of survival and a logistic regression model for the 12 month mortality endpoint were optimized using forward stepwise selection (p<0.05).
Result:
Median follow-up time was 80.2 and 20.2 months in dataset 1 and 2, respectively. MHD (HR=1.023, p=0.001) and WHO performance status (HR=1.25, p=0.03) were selected in the Cox model for dataset 1. Tumor volume (HR=1.0015, p=0.001), WHO performance status (HR=1.023, p=0.02) and MHD (HR=1.0030, p=0.03) were selected in dataset 2. Adding time-dependent covariates revealed a decreasing GTV HR over time in dataset 1 (p=0.02), while MHD risk did not significantly change with time. Worse survival observed in a high MHD subgroup indeed only starts after 8 months (Figure 1). 12 month survival modeling included the covariates MHD (optimal cut-off 22Gy) and GTV (AUC=0.71). In dataset 2, these covariates and cut-off resulted in a model with AUC=0.63. Figure 1
Conclusion:
Mean heart dose is an independent risk factor for early mortality in two cohorts with different treatment periods and techniques. The best classifier for 12 month mortality risk was obtained with the MHD<22Gy constraint, which could be used in model-based implementation of proton therapy.
