Click here to

Session: HyTEC SBRT Dose Constraints [Return to Session]

HyTEC SBRT Dose Constraints

J Grimm1*, E Yorke2*, A Mahadevan3*, A Jackson4*, L Marks5*, (1) Geisinger Health System, Wilkes Barre, PA, (2) Memorial Sloan-Kettering Cancer Center, New York, NY, (3) Geisinger Medical Center, Danivlle, PA, (4) Memorial Sloan-Kettering Cancer Center, New York, NY, (5) University of North Carolina at Chapel Hill, Chapel Hill, NC


10:30 AM HyTEC Overview - J Grimm, Presenting Author
10:41 AM HyTEC - The Project and the Product - E Yorke, Presenting Author
10:52 AM Pancreas TCP as a Case Study of HyTEC Process - A Mahadevan, Presenting Author
11:03 AM Physicist/Modeler's Perspective on HyTEC - A Jackson, Presenting Author
11:14 AM Physician Perspective on HyTEC - the Clinical Needs, Opportunities and Limits - L Marks, Presenting Author

TU-A-TRACK 7-0 (Tuesday, 7/27/2021) 10:30 AM - 11:30 AM [Eastern Time (GMT-4)]

HyTEC Overview: The HyTEC (High Dose per Fraction, Hypofractionated Treatment Effects in the Clinic) review is an effort to summarize the currently-available stereotactic body radiation therapy (SBRT) dose / volume / outcome data to update and refine the normal tissue tolerance guidelines provided by the classic “Emami” paper (IJROBP May 1991) and the recent QUANTEC publications (IJROBP March 2010), for radiosurgery throughout the body in 1 to 5 fractions. HyTEC was formed as the SBRT Working Group (WGSBRT) within the American Association of Physicists in Medicine (AAPM) under the Biological Effects Subcommittee (BESC) and consists of more than 100 authors who are physicians, physicists, radiobiologists, and biomathematicians. The effort has resulted in 7 articles addressing NTCP, 9 articles addressing TCP, and several vision and editorial articles, that will constitute a special issue of the Red Journal (May or June 2021) with free access for AAPM members. This session will provide an updated summary of the overall effort, with an emphasis on general principles and findings of the project. The recently completed Pancreas TCP paper is an example of the methods and results.

HyTEC_ the project and the product: The 16 HyTEC organ/tumor-specific papers are analyses of selected, peer-reviewed publications of sbrt treatments spanning the entire body: Cranial, Head and Neck, Thoracic, Abdominal, Pelvic, and Spinal. Results are presented in a format similar to that of QUANTEC. For each paper, a self-assembled group of experts (physicians, physicists, radiobiologists and biomathematicians) systematically triaged the literature to find papers that included sufficient data for a quantitative synthesis, preferably (but not always) in the form of a sigmoidal curve -a mathematical model-usually requiring a radiobiological dose correction to account for fractionation. Additionally, each paper summarizes such site-specific questions as volume definition and ambiguities in outcome determination. From the models, readers can determine dosimetric metrics and uncertainties associated with literature-based NTCP or TCP outcomes and can compare them with their own experience. As in QUANTEC, the HyTEC introductory article summarizes consensus dose and dose-volume metrics for the major SBRT treatment types, risk organs and tumors in tabular form, such that readers can evaluate and compare them with those used in their own clinics.

Pancreas TCP: Pancreatic cancer is the fourth leading cause of cancer death in the United States, though it is only 3% of all new cancers. Five-year overall survival (OS) is less than 25%. Treatments often combine chemotherapy, surgery and radiation therapy with a major radiation therapy goal of converting locally advanced and borderline resectable pancreatic cancer (LAPC and BRPC) to resectable. SBRT is a desirable form of radiation treatment because short treatment schedules interfere minimally with chemotherapy. However, treatment planning and delivery are complicated by respiratory motion and inter- and intra-fraction normal organ variability and proximity of the tumor to stomach and bowel limits safe prescription doses.
A PubMed search initiated in 6/2019 resulted in 414 articles of which only 32 met the inclusion criteria of providing the cumulative prescription, number of fractions, number of patients, Kaplan-Meier (KM) or individual patient data on OS or Local Control (LC), at least 75% of patients without prior pancreatic radiation and at least 90% without distant disease. These articles provided 48 data points; all but three studies used between 1 and 5 fractions. Study sizes ranged from 12 to 91 patients. There was considerable variability among these studies including the starting points for effect duration (e.g. time from diagnosis vs time from start of radiation), different applications of respiratory motion control and image guidance, the definition of local failure. Detailed dose distributions were rarely provided. For study comparisons, prescription doses were converted to equivalent dose in 3-fractions using the Linear-Quadratic model with α/β=10 Gy (3fxED). Given the variability, modeling LC beyond a year was not attempted.
Resectability was the strongest factor affecting outcome; average 1 year LC was over 90% for patients with R0 resection. For unresected patients, 1 year LC was <70% for 3fxED<24 Gy, approximately 77% for 3fxED approximately 28.2 Gy and approximately 86% for 3fxED of 36 Gy. The dose response from studies in which at least 80% of patients were unresected throughout the study was significant; a logistic model was constructed: 1-year LC=100%/(1+(D50/D)4 g50 ) where D50=17.6 Gy (95% CI 8.8-21.5) and g50=0.64 (95% CI 0.27-1.02).
It is desirable for future studies to agree on starting time for KM analysis, to better stratify patients by extent of resectability, provide more detail about motion management and image guidance and provide dosimetric information beyond the prescription dose.

Physicist/Modeler’s perspective: The HyTEC project has used published literature to investigate and synthesize normal tissue complication probabilities in 7 tissues and local control in 9 tumor types resulting from hypo-fractionated and single fraction radiotherapy/SBRT, resulting in models of treatment tolerances and outcomes.
In this talk we will discuss the successes and limitations of this effort, in terms of the kinds of information that could be gleaned from the published data, and the subsequent uncertainties in the outcomes models, tolerances, and rates of local control. The reporting standards required to overcome the limitations that have hampered HyTEC will be summarized Finally, we will highlight the opportunities for modelling treatment outcome that may result from SBRT treatments.

Physician’s perspective: Caring for patients with cancer is a privilege. We are fortunate to be able to offer patients and their families the opportunity of cure from an often-lethal disease, as well as to relieve much pain and suffering. At the same time, with responsibility comes a level of burden. Strikingly an appropriate balance between attaining local control/cure vs. minimizing the risk of serious morbidity can be difficult. While we have “class solutions” for common clinical situations, there are often the patient-specific issues (e.g. anatomy, comorbid conditions) that can make treatment planning quite challenging. In our quest to best serve our patients, we try to make data driven objective decisions. However, sorting through the literature can be a daunting task- especially for busy clinicians. Thus, expert reviews that provide readily accessible, clinically-useful information can be most helpful. The dose/volume/outcome data provided by the prior QUANTEC, the current HyTEC, and similar initiatives (e.g. PENTEC; Pediatric Normal Tissue Effects in the Clinic), provide this type of helpful information, and we believe that these efforts have helped, and will continue to help, a large number of patients. At the same time, physicians/planners need to be cognizant of the limitations inherent in these systematic/ pooled reviews. For example, there are inherent inaccuracies in the pooling of data from different centers (e.g. related to differences in image segmentation, treatment techniques, outcome definitions; reporting biases; patient heterogeneities; and mathematical models of uncertain accuracy). In addition, our estimates of local control and toxicity risk may appear overly optimistic; e.g. in some settings local recurrence may be difficult to establish with certainty by non-invasive imaging methods, and/or complications may not be accurately recognized, and/or there is a bias towards the publication of favorable results. Further, a wide-spread problem in much of our literature is the use of actuarial techniques to compute estimates of local control, with the censoring patients at the time of death. The accuracy of actuarial techniques requires that censoring events should be independent of the endpoint under consideration. Since the pace of disease beyond the treated site (that can cause the censoring event of death) and the pace of regrowth of treated site (that obviously impacts local recurrence) are likely related, actuarial estimates may not be accurate and may overstate the local control. Nevertheless, despite these limitations, we believe that the data provided by the HyTEC reviews are useful. These issues, as well as an “Idealized Future State” to yield better data (e.g. leveraging advances in electronic health records; standards for contouring, structure naming, outcome reporting; and patient empowerment) will be discussed.

Learning Objectives:
1. Understand AAPM’s role in the HyTEC effort
2. Understand some ways in which a clinical physicist or dosimetrist can use the HyTEC ‘organ papers’
3. Understand the efficacy of current SBRT techniques for treatment of pancreatic cancer
4. Understand causes of successes and failures of outcome modeling for SBRT
5. Understand future opportunities in building on reported clinical experience to improve SBRT

Funding Support, Disclosures, and Conflict of Interest: JG: patent for DVH evaluator; grants from Accuray, Novocure AM: Varian and Accuray Advisory Board EY, AJ: NCI P30CA008748



Not Applicable / None Entered.


Not Applicable / None Entered.

Contact Email