Expert Insights: GI Cancer
Trends in Proton and Carbon Ion Therapy to Assist Clinical Choices and Research Pursuits
Dosimetric Predictors of Radiation-Induced Liver Disease with the Use of Proton Beam Therapy for Hepatocellular Carcinoma
Smith "Jim" Apisarnthanarax, MD
University of Washington, Seattle Cancer Care Alliance
My last gastrointestinal Expert Insight review (below) focused on hepatocellular carcinoma (HCC) and highlighted some of the technical aspects of proton beam therapy (PBT). This review continues the spotlight on HCC by discussing a recent article that investigated dosimetric predictors of radiation-induced liver disease (RILD) that are specific to PBT . Historically in the photon stereotactic body radiation therapy (SBRT) literature, recommendations for normal liver (liver minus gross tumor volume) constraints for the cirrhotic patient population commonly included the mean liver dose as a significant predictor of RILD (most commonly non-classic RILD, as classic RILD is rarely seen in the modern era of liver radiation therapy). For example, both the QUANTEC guidelines and recent data from investigators at the Princess Margaret Hospital (Toronto, ON) recommend limiting the mean liver dose to below 13-16 Gy [2, 3], which is consistent with the normal liver constraints included in the RTOG 1112 and NRG GI-003 randomized trials.
However, recent data from Hsieh et al. suggest that mean liver dose may not be the most relevant dosimetric variable to consider for HCC patients treated with PBT . Investigators of this multi-institutional retrospective study of 136 patients from one Eastern country (Taiwan, n = 102) and one Western country (U.S., n = 34) analyzed the correlation of various dosimetric variables with classic and non-classic RILD and introduced two relatively novel concepts. The first concept was defining the unirradiated liver volume (ULV) or spared liver as the normal liver receiving less than 1 Gy. This definition of “spared liver” contrasts with previous SBRT studies that have recommended 15-18 Gy as the doses to limit the critical volume of liver to be spared [3, 4]. The second concept was related to the notion that, in addition to considering the amount of ULV, the metabolic need of the liver relative to s patient’s body size should also be factored into predicting RILD. In other words, the ULV should be normalized according to the standard liver volume (SLV) based on the body surface area of patients with the ULV/SLV ratio.
Two key findings from the study are worth discussing. The authors found that the incidence of RILD, including unrecoverable and fatal RILD, significantly increased with decreasing ULV/SLV. In patients with Child-Pugh A cirrhosis, fewer than 10% of patients developed RILD when ULV/SLV exceeded 40%; however, when the ULV/SLV rate was below 30%, the rate of RILD was nearly 40% and fatal RILD occurred in 6% of patients. The difference was even more striking in Child-Pugh B cirrhotic patients: if ULV/SLV exceeded 40%, only 14% of patients developed RILD versus 83% in those with ULV/SLV below 40%. The second key finding was that on multivariable analysis ULV/SLV was the only dosimetric variable that independently predicted for RILD. ULV alone (without normalizing to SLV), liver V1 (volume of liver receiving 1 Gy or higher), and mean liver dose did not independently predict for RILD.
Overall, these data may support a paradigm shift in how treatment planning is approached for HCC patients treated with PBT. Although these data will need to be reproduced by other studies, they are thought-provoking in that they suggest PBT planning may require different dosimetric constraints compared to photon-based planning. This idea may not seem so far-fetched when considering the difference in physical characteristics between protons and photons. With protons, the mean liver dose is primarily driven by the entrance dose to the normal liver and the dose to the normal liver that overlaps with the planning target volume (PTV); the dose to the normal liver outside the PTV or Bragg peak is typically nominal. The mean liver dose in photon SBRT-treated livers, however, is related to both the high dose within the normal liver and PTV overlap as well as the moderate and low dose bath to the normal liver outside of the PTV. The mean liver dose, therefore, does not have the same dosimetric implications for protons as it does for photons and very well may explain why the clinical significance of the mean liver dose is less critical for PBT. In the surgical world, the volume of liver spared (not resected) is a key factor in determining postoperative liver failure. The Bragg peak of protons is perhaps most akin to the cutting edge of a surgeon. This conceptual similarity may also help explain why the ULV (essentially receiving nearly no dose of radiation) is most relevant for PBT. How important it is to consider the metabolic needs of the patient in relation to ULV is less clear, particularly since the statistical significance of ULV/SLV appeared to be different between the Eastern and Western cohorts.
1. Hsieh CE, Venkatesulu BP, Lee CH, Hung SP, Wong PF, Aithala SP, Kim BK, Rao A, Tung-Chieh Chang J, Tsang NM, Wang CC, Lee CC, Lin CC, Tseng JH, Chou WC, Wang YC, Krishnan S, Hong JH. Predictors of Radiation-Induced Liver Disease in Eastern and Western Patients With Hepatocellular Carcinoma Undergoing Proton Beam Therapy. Int J Radiat Oncol Biol Phys. 2019;105:73-86. doi: 10.1016/j.ijrobp.2019.02.032. Epub 2019 Feb 21. PMID: 30797890.
2. Pan CC, Kavanagh BD, Dawson LA, Li XA, Das SK, Miften M, Ten Haken RK. Radiation-associated liver injury. Int J Radiat Oncol Biol Phys. 2010;76:S94-100. doi: 10.1016/j.ijrobp.2009.06.092. Review. PMID: 20171524.
3. Velec M, Haddad CR, Craig T, Wang L, Lindsay P, Brierley J, Brade A, Ringash J, Wong R, Kim J, Dawson LA. Predictors of Liver Toxicity Following Stereotactic Body Radiation Therapy for Hepatocellular Carcinoma. Br J Cancer. 1988;57:313-6. PMID: 2833301. PubMed Central PMCID: PMC2246512. doi: 10.1038/bjc.1988.69.
4. Son SH, Choi BO, Ryu MR, Kang YN, Jang JS, Bae SH, Yoon SK, Choi IB, Kang KM, Jang HS. Stereotactic body radiotherapy for patients with unresectable primary hepatocellular carcinoma: dose-volumetric parameters predicting the hepatic complication. Int J Radiat Oncol Biol Phys. 2010;78:1073-80. doi: 10.1016/j.ijrobp.2009.09.009. Epub 2010 Mar 6. PMID: 20207492.
Technical Aspects in the Use of Proton Beam Therapy for Hepatocellular Carcinoma
Within gastrointestinal malignancies, hepatocellular carcinoma (HCC) is perhaps one of the least controversial and more universally accepted indications for the application of proton beam therapy (PBT). Many HCC patients may be potentially cured with the appropriate dose of radiation and the stakes are high for harm; radiation-related hepatotoxicity is one of the rare radiation-related complications that can be life-threatening. HCC is one of the disease sites that the ASTRO PBT Model Policy supports as medically necessary and justified according to clinical data . The NCCN Guidelines, version 2.2019, also support the use of PBT when clinically appropriate . Several studies on PBT for HCC have recently been or will soon be published, and will be the focus of upcoming Expert Insight reviews. This review will direct attention to some of the technical aspects of PBT for liver cancers, which can be quite complex. Guidelines and standardization of technical standards regarding PBT for HCC are limited; the recently released ACR-ASTRO Practice Parameter for the Performance of Proton Beam Radiation Therapy is a recent example .
A recent study from the Samsung Medical Center (Seoul, South Korea) reported the use of PBT in 101 patients with HCC with predominantly well-compensated liver function (89% Child-Pugh A) to a range of doses, most commonly with either 66 GyE in 10 fractions (92 EQD2a/ß10) or 50 GyE in 10 fractions (62 EQD2a/ß10) as consolidative (primarily after TACE) or salvage treatment in 79% of patients . Both uniform scanning and pencil-beam scanning techniques were used; the latter was only used in the breath-hold setting. The median follow-up was relatively short at 4.9 months. At 3 months, the local control rate, defined by at least stable disease radiographically, was high at 97%. Complete radiographic response was observed in 69% of patients. PBT-related hepatotoxicity was low with a Child-Pugh score increase of 2 or more reported in only 4% of patients. PBT-related gastroduodenal toxicity was similarly low at 5%.
Although the number of patients and follow-up period in this study are relatively low in comparison to other published HCC PBT studies from Asia, the details of the simulation and image guidance from this study are well-described and worth discussion. Motion management was employed in 84% of cases with breath-hold being used the majority (73%) of the time. For simulation and treatment delivery, an in-house biofeedback system was used in which the patients’ breathing tracings as a function of time and amplitude were visible to patients via liquid crystal display (LCD) glasses. For breath-hold cases, a visual prompt was displayed on the LCD glasses screen to notify patients when to hold their breath. For voluntary shallow breathing, a visual display of the upper inhalation and lower exhalation lines guided patients on how to breathe. During simulation, 6 CT scans were obtained: 2 non-IV-contrast scans, 2 IV-contrast phase (arterial, venous) scans, and 2 additional scans to check for inter-fractional motion variation. For those patients undergoing respiratory-gated and/or free-breathing methods, 4-dimensional CT was obtained. The authors noted that cone-beam CT for image guidance was used along with orthogonal films (it was unclear which patients received which image guidance).
The importance of accounting for and managing motion in liver cancer patient treated with PBT cannot be under-stated. Aside from minimizing the well-known uncertainties that come along with PBT planning and delivery, motion management for liver cases will also minimize the amount of normal liver irradiated, which is critical for HCC patients with cirrhosis. In the setting of breath-hold techniques, the maximal minimization of target volume occurs from the elimination of needing an internal target volume expansion from the gross tumor volume, resulting in maximal reduction in the irradiated normal liver volume. Respiratory-gating is also an attractive alternative. However, the availability of true automated respiratory-gating is currently limited in the proton community. When breath-hold or respiratory-gating is infeasible or inappropriate, then 4-dimensional CT imaging during simulation is essential to estimate tumor and organ-at-risk motion for contouring and treatment planning . As was performed in the Yu et al. study , multiple CT scans were obtained during simulation and included multi-phase IV contrast imaging. Recent data have shown that multi-phase contrast imaging is important for accurate delineation of tumor targets for HCC and intrahepatic cholangiocarcinoma patients . Last but not least is image-guidance. Many proton centers only have orthogonal KV imaging available for image-guidance. Similar to the unease and hesitancy of delivering photon-based SBRT without 3-dimensional volumetric image guidance such as cone-beam CT or CT-on-rails, many physicians are hesitant to employ stereotactic body radiation therapy regimens with PBT in HCC patients. On-board volumetric imaging, however, is increasingly becoming available in newer proton facilities; the increased use of hypofractionated (e.g. SBRT-PBT) regimens for HCC should then be expected.
4. Yu JI, Yoo GS, Cho S, Jung SH, Han Y, Park S, Lee B, Kang W, Sinn DH, Paik YH, Gwak GY, Choi MS, Lee JH, Koh KC, Paik SW, Park HC. Initial clinical outcomes of proton beam radiotherapy for hepatocellular carcinoma. Radiat Oncol J. 2018 Mar;36(1):25-34.
5. Niska JR, Keane FK, Wolfgang JA, Hahn PF, Wo JY, Zhu AX, Hong TS. Impact of intravenous contrast enhancement phase on target definition for hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (IHC): Observations from patients enrolled on a prospective phase 2 trial. Pract Radiat Oncol. 2016 Jan-Feb;6(1):e9-16. doi: 10.1016/j.prro.2015.08.005