468 / 1082
S453
ESTRO 36 2017
_______________________________________________________________________________________________
Conclusion
Auto-Plan in Pinnacle allowed fast planning of FFF VMAT
plans for partial breast cancer radiotherapy. Compared to
forward planned tangential radiotherapy the WMAT plans
were better at both sparring of the ipsilateral lung and in
covering the PTV. The VMAT plans could be delivered
quickly, and as a result patients treated in breath hold
could be treated with half the number of breath holds.
Poster: Physics track: (Radio)biological modelling
PO-0846 Bowel dose-volume relationship for patient-
reported acute intestinal toxicity from pelvic IMRT
C. Sini
1
, B. Noris Chiorda
2
, P. Gabriele
3
, G. Sanguineti
4
, S.
Morlino
5
, F. Badenchini
6
, D. Cante
7
, V. Carillo
8
, M.
Gaetano
8
, T. Giandini
9
, V. Landoni
10
, A. Maggio
11
, L.
Perna
1
, E. Petrucci
7
, V. Sacco
2
, R. Valdagni
12
, T. Rancati
6
,
C. Fiorino
1
, C. Cozzarini
2
1
San Raffaele Scientific Institute, Medical Physics,
Milano, Italy
2
San Raffaele Scientific Institute, Radiotherapy, Milano,
Italy
3
IRCC-Candiolo, Radiotherapy, Torino, Italy
4
Regina Elena Institute – IFO, Radiotherapy, Roma, Italy
5
Fondazione IRCCS Istituto Nazionale Tumori, Radiation
Oncology, Milano, Italy
6
Fondazione IRCCS Istituto Nazionale Tumori, Prostate
Cancer Program, Milano, Italy
7
ASL TO4 Ospedale di Ivrea, Radiotherapy, Ivrea, Italy
8
Centro AKTIS Diagnostica e terapia, Radiotherapy,
Napoli, Italy
9
Fondazione IRCCS Istituto Nazionale Tumori, Medical
Physics, Milano, Italy
10
Regina Elena Institute – IFO, Medical Physics, Roma,
Italy
11
IRCC-Candiolo, Medical Physics, Torino, Italy
12
Fondazione IRCCS Istituto Nazionale Tumori, Radiation
Oncology- Prostate Cancer Program- UNIV Hematology
and Hemato-Oncology- Università degli Studi di Milano,
Milano, Italy
Purpose or Objective
Intestinal toxicity (IT) may affect the quality of life of
patients (pts) treated with whole-pelvis intensity-
modulated radiotherapy (WPIMRT) for prostate cancer.
The aim of this investigation is to identify quantitative
bowel dose-volume relationships for acute patient-
reported IT.
Material and Methods
A cohort of pts was enrolled in 6 Institutions within a
registered prospective trial. Pts were treated with
conventional or moderate hypo-fractionation to
prostate/prostatic bed and WPIMRT delivering 51.8 Gy
(median dose, range: 50.4-56.1 Gy) to pelvic nodes while
sparing the bowel outside PTV as much as possible. Acute
IT was evaluated by the Inflammatory Bowel Disease
Questionnaire pertaining to the Bowel Domain (IBDQ-B)
filled in by pts at baseline and at mid-point/end of RT.
IBDQ-B includes 10 items (#item) scored on a 7-point scale
(worst symptoms=lower scores).
The 25
th
percentiles of the most severe worsening (Δ)
between baseline and half/end RT were set as end-points.
The correlation between end-points and bowel loops cc/%
DVH/DSH (from V5Gy to V60Gy) as well as selected clinical
parameters was investigated through multi-variable
logistic regression. Goodness of fit was estimated by the
Hosmer Lemeshow test (HL) and the Brier score (BS);
performances of the model were assessed by the
calibration plot. Internal validation was performed by
1000 bootstrap resampling.
Results
Data of 206 pts (80 radical, 79 adjuvant, 47 salvage RT)
were available: 25/109/72 pts were treated with fixed-
fields, rotational and Tomotherapy technique
respectively. A relatively small but significant Δ (p<0.05)
was found for all questions: the median Δ was 2 points for
#1 (bowel movements) and 1 point for #5 (loose stools),
#17 (gas passage) and #24 (urgency to defecate with
empty intestine); Δ was 0 for the remaining 6 items that
were then disregarded in current analysis. No DVH/DSH
parameters were correlated with Δ, except for ΔIBDQ5≤-3
(25
th
percentile, 43/191).
The resulting model after backward selection of variables
(R
2
=0.89, slope:1.037, optimism corrected BS=0.17, Figure
1) included absolute V42Gy and age (protective). Due to
the correlation between DVH variables, three values
representing ‘low’ (V20), ‘intermediate’ (V30) and ‘high’
(V42) dose levels were also considered to define an overall
'DVH-shape” predictor.
When grouping pts according to best cut-off values
assessed by ROC curves (high risk: V20>470cc, V30>245cc,
V42>110cc; low risk: the other pts, figure 2), an
alternative model including high-risk DVH-shape (OR:9.3)
and age (protective, OR:0.94) may be suggested. The
model showed very good calibration (slope:1.003, R
2
=0.92)
and accurate prediction even after bootstrap-based
internal validation (corrected BS=0.16).