ESTRO 36 Abstract Book

657 / 1082

S641 ESTRO 36 2017
_______________________________________________________________________________________________
Gy photon OR electron boost to the tumor bed) were
evaluated with special focus on documented skin toxicity
during RT course.
Acute skin erythema (AE) was visually assessed and
recorded using the RTOG scoring system, before RT and
every 5 fractions. In this study, grade 2-3 AE during RT was
considered as the primary endpoint.
A number of relevant clinical risk factors was
prospectively recorded: age, skin phototype, smoking
habits, use of drugs, neoadjuvant chemotherapy with
anthracyclines and/or taxanes and/or trastuzumab,
hormone therapy with tamoxifen or aromatase inhibitors,
comorbidities and related drugs, T stage, location of
breast surgery.
Dosimetric feature were extracted from the skin dose-
volume histogram for the whole treatment (DVH, absolute
volume in cc), with skin defined as the difference between
the body contour and a 5mm inner isotropic contour from
the body.
Dosimetric and clinical variables were included into
multivariable logistic regression. Goodness-of-fit was
evaluated through Hosmer-Lemeshow test (HL) and
calibration plot.
Results
a total of 147 breast cancer patients (median age 55 years,
range 34–77) were enrolled.
Grade 1, 2 and 3 AE were 65/147 (44%), 52/147 (35%) and
24/147 (16%), respectively.
At univariate analysis only the dose to 20 cc of breast and
use of aromatase inhibitors vs tamoxifen resulted as
predictive factors for toxicity (61.8% vs 17.9%, p>0.01, for
aromatase inhibitors vs tamoxifen, respectively).
ML resulted in a two variable model including the dose to
20 cc of skin (continuous variable, OR=1.09, 10th-90th
percentile 1-1.19) and use of aromatase inhibitors
(OR=1.7, 10th-90th percentile 1.1-2.7). Calibration was
good (HL test p=0.35, calibration slope 1.08). Results for
model and calibration are presented in the figure.
Smoking also resulted to be a risk factor (OR=4) in a
reduced population (87 pts), it was not directly inserted
into ML model due to the high prevalence of missing
values, but it deserves attention and further analysis
Conclusion
this analysis shows that moderate/severe acute skin
erythema is related to skin DVH, particularly to the dose
to 20cc of skin. In the frame of the here used skin
definition, this approximately corresponds to an area of
6x6 cm^2. Use of aromatase inhibitors acts as a dose
sensitizing factor for kind of toxicity.
EP-1195 Regional nodal recurrences after adjuvant
breast radiotherapy – are we covering the target?
L.E. Beaton1, L. Nica1, K. Sek2, G. Ayers1, C. Speers3, S.
Tyldesley1 , A. Nichol1
1 British Columbia Cancer Agency, Radiation Oncology,
Vancouver, Canada
2 British Columbia Cancer Agency, Radiology, Vancouver,
Canada
3 British Columbia Cancer Agency, Breast Outcomes Unit,
Vancouver, Canada
Purpose or Objective
For all breast cancer patients, adjuvant radiotherapy (RT)
reduces locoregional recurrence and for high risk patients,
regional nodal irradiation (RNI) improves overall survival.
However, there is limited data on the anatomical location
of regional nodal recurrence (RNR) after adjuvant RT.
Nodal radiotherapy fields have historically been defined
using anatomical landmarks but with the advent of 3D
radiotherapy planning nodal contouring atlases have been
developed. Validation of these atlases is scarce. Our
objective was to map the location of RNR in patients
previously treated with adjuvant RNI, and assess whether
the treating RT fields provided adequate coverage. We
also assessed whether these areas of RNR were within the
boundaries of the Radiation Therapy Oncology Group
(RTOG) nodal atlas.
Material and Methods
Between 2005 and 2013, we identified 32 patients
previously treated with definitive surgery and adjuvant
RNI for breast cancer that developed RNR detected with
18-fluorodeoxyglucose positron emission tomography
(FDG-PET) imaging, before salvage treatment for RNR.
FDG-PET positive regional lymph nodes were contoured on
each individual PET scan. Deformable registration was
used to fuse the FDG-PET scan with the patient’s original
RT simulation scan, onto which the RTOG atlas had been
retrospectively contoured. Each nodal area of recurrence
was categorized as: in-field, defined as ≥ 95% of the RNR
volume receiving ≥ 95% prescribed dose; marginal, RNR
receiving < 95% prescribed dose; and out of field, RNR not
intentionally covered with the original RT plan. RTOG
coverage was defined for each RNR as ‘inside’, ‘marginal’
or ‘outside’.
Results
Of the 32 patients, 12 (37%) had limited RNR and 20 (63%)
had RNR in addition to distant metastatic di sease on FDG-
PET imaging. 27 (84%) patients received full axillary RT,
3 (9%) supraclavicular fossa (SCF) only, and 14 (44%)
internal mammary node (IMN) RT. Of the 87 nodal
relapses, 17 (20%) were out of field. Of those intentionally
treated, 10 (33%) patients developed SCF relapse, 18 (66%)
axillary relapse and 5 (36%) IMN relapse. 15 (68%) of SCF,
20 (50%) axillary and 1 (14%) IMN nodes were in-field
relapses. The RTOG atlas covered 13 (60%) SCF, 20 (50%)
axilla and 0 (0%) of IMN nodal relapses.