The Prospective SPOTLESS Trial: Setup Accuracy of Tattoo-Less Surface-Guided Breast Radiotherapy Including Regional Nodal Irradiation
, Sophia Albert 1,2, Bahar Cepni 1,2,3, David Neugebauer 1,2, Hin Hoi Lau 1,2, Julian Thater 1,2,3, Klaus Herfarth 1,2,3,4, Stephan Mende 1,2,3, Fabian Weykamp 1,2,3,4,5, Adriana Ayestaran-Aldaz 1,2,3,4, Line Hoeltgen 1,2,3, Nathalie Arians 1,2,3, Jakob Liermann 1,2,3,4,5, Lars Wessel 1,2,3, Semi Harrabi 1,2,3,4, Hanna Waldsperger 1,2,3,4, Jürgen Debus 1,2,3,4,5, Sebastian Klüter 1,2 and Vania Batista 1,2Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe authors make an important contribution to the field of radiation oncology. There is an increasing need to transition radiation therapy daily positioning to non-tattoo-based systems. While there has been literature on this topic for patients being treated to intact breast, the authors have placed emphasis on the treatment of patients requiring regional nodal radiation therapy which is an under studied population in radiation oncology. The paper is sound and well written. The description of the problem is excellent and the methods chosen to validate the hypothesis are excellent and well presented. The results are clear with sound statistical support. The discussion is thorough and identifies both strengths and weaknesses of the analysis including a good description of the challenges in defining the internal mammary lymph node region as a volume. The paper provides confidence that, with experience, this cohort of patients can be treated safely with image guidance with this paper providing strong evidence supporting the thesis.
Author Response
The authors make an important contribution to the field of radiation oncology. There is an increasing need to transition radiation therapy daily positioning to non-tattoo-based systems. While there has been literature on this topic for patients being treated to intact breast, the authors have placed emphasis on the treatment of patients requiring regional nodal radiation therapy which is an under studied population in radiation oncology. The paper is sound and well written. The description of the problem is excellent and the methods chosen to validate the hypothesis are excellent and well presented. The results are clear with sound statistical support. The discussion is thorough and identifies both strengths and weaknesses of the analysis including a good description of the challenges in defining the internal mammary lymph node region as a volume. The paper provides confidence that, with experience, this cohort of patients can be treated safely with image guidance with this paper providing strong evidence supporting the thesis.
>> We thank the reviewer for their time and careful evaluation of our manuscript. We would like to point out that weh ave corrected language and grammatical errors throughout the manuscript and included an additional figure. All changes made to the text are highlighted in yellow.
Reviewer 2 Report
Comments and Suggestions for AuthorsThis is a well-conducted prospective trial addressing a genuine gap: most SGRT setup data come from whole-breast or partial-breast irradiation, whereas regional nodal irradiation is far less documented. The per-level manual grading combined with forward dose recalculation on corrected CBCT is a real strength. The manuscript is clearly written, but several numerical inconsistencies need correction.
Main points:
1.Numerical error in Section 3.3 (lines 242–244). For Level 3/4 (n = 420 fractions), the text states good-to-minor deviations in 95.0% (n = 420) and major deviations in 5.0% (n = 22). These figures do not add up: 95.0% of 420 is 399, and 22/420 is 5.2%. Please recheck and correct.
2.Directional labels appear scrambled (lines 205–208). The sentence reports “lateral, longitudinal, and vertical” deviations but then lists “superior/anterior”, “right/left” and “anterior/posterior”. Lateral corresponds to right/left, longitudinal to superior/inferior, and vertical to anterior/posterior. Also, “superior/anterior” is not a valid axis — presumably superior/inferior was intended.
3.Table 2 is internally inconsistent. The overall median target CTV is given as 641 cm³ (155–1690), yet the medians for all three subgroups are higher (chest wall 792, breast 1278, implant 1062) and their minima (505, 895, 553) are all above 155. In addition, the median PTV (1136 cm³) is smaller than the median breast CTV (1278 cm³), which is not possible with a 5–7 mm expansion. Please verify these values.
4.Clustering of observations is not accounted for. The 457 data points come from only 30 patients, so fractions within a patient are not independent. Chi-square, Fisher’s exact test and standard logistic regression assume independence. A mixed-effects model or GEE with patient as a random effect would be more appropriate, or at least this limitation should be stated explicitly.
5.Definition of the dosimetric coverage metric. “Median CTV dose coverage of 100.0% of the original target dose” is not defined — is this D95%, V95%, mean dose, or D98%? Please specify. Also, the lower end of the range for the internal mammary CTV (57.1%) is striking and deserves comment: how many fractions fell below a clinically acceptable threshold, and were these concentrated in individual patients?
6.Rotational data are not reported. The institutional tolerance included 3° rotation, and the Discussion attributes the internal mammary chain findings partly to rotational errors, yet no rotational results are presented. If CBCT registration allowed 6-degree-of-freedom assessment, please report pitch/roll/yaw; if only translations were evaluated, please state this as a limitation.
7.Reliability of the manual grading. The four-tier scale is subjective, particularly the distinction between “minor” and “major”. Please state how many observers graded the scans, whether they were blinded, and whether inter-observer agreement was assessed. If a single observer performed the grading, this should be acknowledged as a limitation.
8.Denominators for the per-level analyses. The numbers of fractions differ between levels (334, 420, 426 in Section 3.3; 60–86 in Table 3) without explanation. A short sentence noting that not all patients received all nodal levels would help the reader.
9.Conclusion is somewhat strong for the sample size. With 30 patients, the study is not powered to exclude associations between setup accuracy and BMI, breathing technique or target volume. Please soften “irrespective of patient-, tumor-, or treatment-specific variables” — absence of statistical significance is not evidence of absence of an effect. A 95% confidence interval for the 97.6% primary endpoint would also strengthen the results.
Author Response
This is a well-conducted prospective trial addressing a genuine gap: most SGRT setup data come from whole-breast or partial-breast irradiation, whereas regional nodal irradiation is far less documented. The per-level manual grading combined with forward dose recalculation on corrected CBCT is a real strength. The manuscript is clearly written, but several numerical inconsistencies need correction.
>> We thank the reviewer for their time and careful evaluation of our manuscript. We have revised the text, all changes made to the text are highlighted in yellow. Please find point by point comments below:
Main points:
1.Numerical error in Section 3.3 (lines 242–244). For Level 3/4 (n = 420 fractions), the text states good-to-minor deviations in 95.0% (n = 420) and major deviations in 5.0% (n = 22). These figures do not add up: 95.0% of 420 is 399, and 22/420 is 5.2%. Please recheck and correct.
>> We have corrected the numbers: „Level 3/4 (n=442 fractions): Demonstrated good-to-minor deviations in 95.0% (n=420) of cases, while major deviations were present in 5.0% (n=22).“
2.Directional labels appear scrambled (lines 205–208). The sentence reports “lateral, longitudinal, and vertical” deviations but then lists “superior/anterior”, “right/left” and “anterior/posterior”. Lateral corresponds to right/left, longitudinal to superior/inferior, and vertical to anterior/posterior. Also, “superior/anterior” is not a valid axis — presumably superior/inferior was intended.
>> We have corrected the sentence: „The median longitudinal, lateral, and vertical deviations were 2mm (range: 0 – 12mm) in superior/inferior direction, 2mm (range: 0 – 9 mm) in right/left direction, and 2mm (range: 0 – 10 mm) in anterior/posterior direction.“
3.Table 2 is internally inconsistent. The overall median target CTV is given as 641 cm³ (155–1690), yet the medians for all three subgroups are higher (chest wall 792, breast 1278, implant 1062) and their minima (505, 895, 553) are all above 155. In addition, the median PTV (1136 cm³) is smaller than the median breast CTV (1278 cm³), which is not possible with a 5–7 mm expansion. Please verify these values.
>> We appreciate this insightful observation. The table has been thoroughly revised. Specifically, the median values for the subgroups in the original version reflected the PTV instead of the CTV. We have corrected these values in the revised manuscript.
4.Clustering of observations is not accounted for. The 457 data points come from only 30 patients, so fractions within a patient are not independent. Chi-square, Fisher’s exact test and standard logistic regression assume independence. A mixed-effects model or GEE with patient as a random effect would be more appropriate, or at least this limitation should be stated explicitly.
>> We appreciate this insightful observation and have added it to the limitation section: „Another limitation of this study is that multiple observations per patient were analyzed without adjusting for intra-individual clustering, as statistical test assuming independ-ence were applied. Consequently, within-patient dependencies were not fully accounted for.“
5.Definition of the dosimetric coverage metric. “Median CTV dose coverage of 100.0% of the original target dose” is not defined — is this D95%, V95%, mean dose, or D98%? Please specify. Also, the lower end of the range for the internal mammary CTV (57.1%) is striking and deserves comment: how many fractions fell below a clinically acceptable threshold, and were these concentrated in individual patients?
>> We have clarified the exact metric used throughout the manuscript. Target coverage was evaluated using D95% (the dose delivered to 95% of the Clinical Target Volume, expressed as a percentage of the prescribed target dose).
>> We agree that the minimum coverage value of 57.1% warrants detailed clarification. This low value was an isolated extreme outlier occurring in a single fraction, whereas overall coverage across all remaining fractions remained clinically adequate.. We have added this to the results section.
6.Rotational data are not reported. The institutional tolerance included 3° rotation, and the Discussion attributes the internal mammary chain findings partly to rotational errors, yet no rotational results are presented. If CBCT registration allowed 6-degree-of-freedom assessment, please report pitch/roll/yaw; if only translations were evaluated, please state this as a limitation.
>> Than you for raising this point. We have revised the manuscript to clarify that CBCT registration process for both, online guidance and offline analysis, allowed only 3 degrees of freedom (translational: lateral, longitudinal, vertical). As already stated in the methods section (2.3) „rotational correction“ was not allowed. We have revised the methods (2.2 and 2.3) and limitation sections accordingly.
7.Reliability of the manual grading. The four-tier scale is subjective, particularly the distinction between “minor” and “major”. Please state how many observers graded the scans, whether they were blinded, and whether inter-observer agreement was assessed. If a single observer performed the grading, this should be acknowledged as a limitation.
>> A single blinded observer performed all gradings, we have included this to the limitation section: „Finally, all gradings were performed by a single blinded observer. Although this ensured consistent scoring across all data, it precludes the evaluation of inter-observer reliability.“
8.Denominators for the per-level analyses. The numbers of fractions differ between levels (334, 420, 426 in Section 3.3; 60–86 in Table 3) without explanation. A short sentence noting that not all patients received all nodal levels would help the reader.
>> We thank the reviewer for this valuable point. Table 1 outlines the patient characteristics, detailing the specific regional nodal irradiations received bye ach patient. To avoid any misunderstanding, we have explicitly highlighted this in the revised text (section 3.3).: „The extent of regional nodal irradiation varied among the cohort, with not all patients undergoing treatment to every lymph node region. Specific details of the target volumes are provided in Table 1.“
9.Conclusion is somewhat strong for the sample size. With 30 patients, the study is not powered to exclude associations between setup accuracy and BMI, breathing technique or target volume. Please soften “irrespective of patient-, tumor-, or treatment-specific variables” — absence of statistical significance is not evidence of absence of an effect. A 95% confidence interval for the 97.6% primary endpoint would also strengthen the results.
>> We sincerely thank the reviewer for this critical and helpful remark. We fully agree that our study with n = 30 patients was not powered to non-inferiority or to definitively exclude small to moderate effects of patient-, tumor-, or treatment-specific variables (such as BMI, breathing technique, or target volume) on setup accuracy. We have revised the limitations and conclusion sections accordingly. We further have added the 95%-CI for the primary endpoint in the results section.
