Locoregional therapy includes ablation, bland embolization, transcatheter arterial chemoembolization (TACE) with drug-eluting beads, Yttrium-90 trans-arterial radioembolization (TARE), or external beam radiation therapy (EBRT) traditionally using X-rays (photons). The advantage provided by proton beams is that the energy is deposited in a specified depth with a sharp decline in dose beyond the target, known as the Bragg peak phenomenon [12]. This technology has now been used to treat many deep tissue tumors with vulnerable surrounding structures. Three types of delivery methods have evolved through time: the first - passively scattered, second - uniform scanned, and most recently - pencil beam scanned [13]. As the liver has low tolerance to radiation and 5–10% of patients develop radiation-induced liver disease following radiation therapy, the focused energy of proton beam treatment minimizes the dose normal tissue receives and thus decreases the risk of radiation-induced liver disease [14].
Researchers at the University of Tsukuba have published their experience with PBT in HCC, having started treatment in 1985 [15, 16]. Several doses and fractionations were studied with preferred regimens depending on the location of the target. Overall 5-year survival rate was 58%, with a 5-year local control rate of 81%. Of the 266 patients, there were three reported grade 3 toxicities and three grade 2 toxicities, such as perforation, bleeding or inflammation of the digestive tract [16]. Studies have also been published to evaluate patient selection for PBT over stereotactic body radiation therapy (SBRT): retrospective data from the Massachusetts General Hospital including 84 patients in photon therapy and 49 patients in PBT showed improved overall survival of 31 months in the PBT group compared to 14 months in the photon therapy group, also demonstrating decreased risk of non-classic radiation-induced liver disease (RILD) (OR = 0.26) [17]. In a dosimetric study comparing different tumor sizes, the average risk of RILD with spot-scanning PBT was 6.2% compared to 94.5% with intensity-modulated radiation therapy in tumors with a gross tumor volume (GTV) nominal diameter more than 6.3cm [18]. Currently, the ongoing NRG Oncology GI-003 trial aims to compare the local control and liver toxicity rates between PBT and photon therapy [19].
In iCCA, EBRT as an adjunct to systemic therapy has shown improved local control and overall survival [20]. Multiple studies have reported the use of PBT at various dosing regimens in iCCA with favorable local control, overall survival, and minimal toxicity (reported biliary complications included cholangitis, sometimes requiring stent placement, and common bile duct stenosis) [21, 22].
The ALBI score was originally developed as a tool for liver function assessment in patients with HCC by Johnson et al. in 2015 [23] and has been validated in several other studies [24, 25]. The score was then applied to cholangiocarcinoma, and in multiple studies for iCCA demonstrated association with both short- and long-term outcomes following resection and microwave ablation [26, 27]. The MELD-Na score was originally developed to predict survival following transjugular intrahepatic portostystemic shunt placement, but has since been modified and adopted for use in prognostication of end-stage liver disease and prioritization of liver transplant candidates [28].
Our study adds to the growing literature documenting the safety and efficacy of PBT in treating large liver cancers. In our population of patients with advanced and treatment experienced cancers (45.2% with prior locoregional therapy; 45.5% with prior chemotherapy, 4.7% with resection, 28.9% with suspicion for vascular invasion, 23.1% with lymph node involvement and 20.5% with extrahepatic metastasis), radiologic response rates were excellent, at 76% and 79% for HCC and iCCA, respectively. In the HCC group, higher pre-treatment albumin levels, higher pre-treatment AFP, BED ≥ 80.5 Gy, and higher total radiation dose all correlated with tumor response, suggesting that “healthier” patients who could tolerate higher doses of treatment, even those with markedly elevated tumor marker, fared better. Similarly, patients with iCCA with higher pre-treatment platelet counts had improved survival. Interestingly, similar to the HCC group, higher pre-PBT CA 19 − 9 correlated with improved tumor response. The reason for this phenomenon is unclear but may suggest a particular sensitivity of tumor-marker producing cancers to radiation therapy.
Liver toxicity is clearly an important consideration in planning EBRT. In our cohort, using MELD-Na and ALBI grade, liver function was shown to be preserved even in those patients with baseline cirrhosis who received BED ≥ 80.5 Gy). This is the threshold which has previously been associated with an optimal balance between tumor control and toxicity [20]. This lends support to the idea that PBT delivers effective therapy to the targeted neoplastic tissue, with minimal toxicity. In the group with HCC, however, higher mean normal liver dose and higher V10Gy was associated with worsening of ALBI grade, confirming that there is an upper ceiling on radiation dose that can be safely used in those with underlying cirrhosis (77% with HCC had confirmed underlying cirrhosis in our cohort).
In a group of HCC patients undergoing SBRT including 101 CTP class A and 13 CTP class B subjects, Velec et al. demonstrated that a higher baseline CTP score, lower platelet count, higher effective liver volume irradiated, and higher dose to 800cc of liver were associated with increased risk of liver function decline [29]. Conversely, Hsieh et al. demonstrated that in 136 patients treated with PBT, gross tumor volume and CTP class were independent RILD risk factors, while mean liver dose and target-delivered dose were not associated with RILD [30]. Given its favorable toxicity profile, PBT is therefore a promising emerging option for the treatment of unresectable HCC and iCCA, and a viable alternative to traditional photon-based EBRT. Specifically, limited data on PBS consists of two studies. The first demonstrated that liver function was minimally affected, with only 2.7% demonstrating significant increase of ALBI grade. Excellent outcomes were achieved, with one-year local control of 94%, progression free survival 35%, and overall survival 78% [10]. The second study by Yoo et al. did not include information on liver toxicities as defined by ALBI grade change or MELD-Na change [11]. Within our HCC group, 77% had cirrhosis and 16.7% had worsening of ALBI grade. Even though 58.3% of our patients demonstrated a decline in liver function in the form of worsening MELD-Na, a third of our cohort (33.3%) actually showed improvement in MELD-Na. In comparing sizes, for hepatocellular carcinoma Bhangoo et al. reported median CTV of 170cc, Yoo et al. reported median maximum diameter of 6.7cm, and Hong et al. reported median GTV of 106.4cc, compared to a median CTV of 144.8cc in our study [10, 24, 11]. In intrahepatic cholangiocarcinoma, Hong et al. reported a median GTV of 133.7cc compared to a median CTV of 280.2cc in our study. Despite the fairly large tumor sizes, 76.2% of our intrahepatic cholangiocarcinoma patients showed radiologic improvement and overall survival was 86.4%, superior to reported data.
Limitations of our study include the non-randomized nature of treatment selection for our patients, as well as lack of a direct comparison group. The retrospective nature of the study also limits certain data collection. Treatment decisions were guided by multidisciplinary discussion and consensus-building and represent a “real-life” experience. The strengths of our study include our cohort which comprised difficult to treat tumors (large tumor size, multifocality), the completeness of data extraction and the relatively long duration of follow-up. Additionally, all our patients received PBT utilizing pencil beam scanning, a relatively new approach. Overall, we were able to demonstrate the applicability and safety of PBT in advanced tumors, including those occurring against a background of cirrhosis. We therefore conclude that PBT with pencil beam scanning for unresectable liver cancers is safe and effective even in difficult to treat tumors.