Virtual Library
Start Your Search
S. Marshall
Author of
-
+
P2.14 - Radiotherapy (ID 715)
- Event: WCLC 2017
- Type: Poster Session with Presenters Present
- Track: Radiotherapy
- Presentations: 1
- Moderators:
- Coordinates: 10/17/2017, 09:30 - 16:00, Exhibit Hall (Hall B + C)
-
+
P2.14-020 - Clinical Validation of NTCP-Models for Esophagus Toxicity in Non-Small Cell Lung Cancer Patients Treated with Concurrent Chemoradiation (ID 9272)
09:30 - 09:30 | Author(s): S. Marshall
- Abstract
Background:
Concurrent chemoradiation (CCRT) is the preferred treatment approach in inoperable non-small cell lung cancer (NSCLC) patients. However, CCRT increases the risk of acute esophagus toxicity (AET) compared to radiotherapy alone or sequential chemoradiation. To minimize the risk of AET, an international RT-dose constraint of either V50Gy or V60Gy <50% is used. However, clinical applicability of those models is not evident since assessment of the model accuracy and validity should be performed before generalizing to other populations. Therefore, the aim of this study was to clinically validate the two NTCP-models to predict acute esophagus toxicity in NSCLC patients treated with CCRT.
Method:
To validate the NTCP-models, clinical data of 274 inoperable NSCLC patients receiving CCRT using IMRT was used. The planned V50Gy and V60Gy and the prospectively scored grade ≥2 AET (CTC-AE) were retrieved and independently reviewed. The grade ≥2 AET probability for the V50Gy and V60Gy was calculated as; [1/1+[exp][[-0.515 + (0.027*V][50]] and [1/1+[exp][[-0.701 + (0.029*V6][0]]]. Validity of the model was assessed with the ability to predict the number of grade ≥2 AET events (calibration) and the ability to distinguish between those who develop grade ≥2 AET from those who do not (discrimination, area under the curve (AUC)). Furthermore, sensitivity and specificity for different cut-off points were determined.
Result:
From the 274 NSCLC patients, 125 (45.6%) patients developed grade ≥2 AET (93.8% grade 2, 6.2% grade 3), median V50Gy 23% (interquartile range 10.1-35.6%), median V60Gy 4.3% (interquartile range 0-20.5%). Calibration showed that the V50 overestimated the risk of developing grade ≥2 AET in low-risk patients while the V60 underestimated the risk of developing grade ≥2 AET in high-risk patients. Discrimination of both algorithms demonstrated a similar moderate fit (AUC 0.70 95%CI 0.64 to 0.76 and AUC 0.69 95%CI 0.63 to 0.76 for the V50Gy and V60Gy, respectively). For V50Gy, a cut-off point of more than 40% probability of developing grade ≥2 AET resulted in the most favorable sensitivity of 95.8% for grade ≥2 and 100% for grade 3, with specificity scores of 54.6% and 40.7% respectively. For V60Gy, a cut-off point of more than 60% resulted in the most favorable sensitivity of 95.1% for grade ≥2 and 100% for grade 3, with remarkably low specificity scores of 9.1% and 18.8%, respectively.
Conclusion:
The NTCP-models to predict acute esophagus toxicity in NSCLC patients both showed good predictive accuracy. For clinical practice, the V50Gy seems to be the most sensitive without compromising safety and efficacy.
