Efficacy
In appropriately selected patients, surgical resection for liver metastases is the standard of care. However, 70-90% of liver metastases are unresectable due to lesion size, location and comorbidities.r Other local treatment options include radiofrequency ablation (RFA), SABR, cryotherapy and selective internal radiation therapy (SIRT). Choice of treatment is dictated by many factors including patient comorbidities, lesion size, lesion number and proximity to critical anatomy such as blood vessels, biliary tract and diaphragm. Median overall survival can range from 10-34 months but is difficult to determine due to the heterogeneity of studies, spectrum of histologies and differences in systemic therapy.r Out of field progression occurs in a substantial proportion of patients. Factors associated with increased survival include hepatic tumour burden, local control of metastases, the absence of extra-hepatic disease, favourable histology and tumours <3 cm.rrrr
Evidence for SABR for liver metastases is confined to retrospective series and prospective phase 1 and 2 trials. There is no prospective comparison of SABR and RFA and attempts to run such a trial have failed due to poor recruitment. There is a significant heterogeneity in primary histology, previous liver therapies, the size and number of lesions treated, dose-fractionation schedule delivered, prescription points and planning criteria. Despite study heterogeneity, local control is closely associated with radiation dose and primary histology.rr
A HyTEC analysis of 13 studies found 1, 2 and 3-year actuarial local control rates for liver metastases after SABR to be 90%, 79% and 76%, respectively. Outcomes were significantly better for lesions treated with biologically equivalent doses exceeding 100 Gy (3-year local control 93%).r
An early phase I-II study by Hoyer et al. treated 64 patients with 141 colorectal metastases with a dose of 45Gy in 3 fractions.r Eligible patients had up to 6 metastases that were unresectable and not amenable for other local treatment. Maximum diameter of the largest metastasis was 6 cm. The majority of patients were treated for liver metastases, 69%, and 33% of patients had previously received local treatment. While 2-year local control (LC) was reasonably high at 79%, 2-year progression-free survival was only 19% with most distant failures occurring in the same organ. 2 and 5-year overall survival (OS) was 38% and 13%, respectively.
Rule et al. established a dose-response relationship for hepatic metastases.r 27 patients were enrolled onto 3 dose escalation cohorts: 30 Gy in 3 fractions, 50 Gy in 5 fractions and 60 Gy in 5 fractions. Eligible patients had 1-5 hepatic metastases, ability to spare a critical hepatic volume (V21 Gy <700 cc), adequate baseline hepatic function, no concurrent anti-cancer therapy, and a Karnofsky performance score of ≥60. 12 (44%) patients had received prior liver-directed therapy to lesions other than those targeted in the study. Median OS was 37 months and median lesion size was 2.5 cm. LC rate was 100% at 2 years for the 60 Gy cohort, 89% for the 50 Gy cohort, and 56% for 30 Gy cohort (Figure 1). Kaplan-Meier analysis revealed a statistically significant difference for LC between the 60 Gy and 30 Gy cohorts, but not between the 60 Gy and 50 Gy cohorts.
Figure 1: Actuarial local control by treatment cohort
© Ann Surg Oncol 2011r
The longest follow up to date has been reported by Scorsetti et al. in which 61 patients with 76 liver metastases of varying histology were treated between February 2010 and September 2011.r Patients had 1-3 liver metastases not suitable for surgery, normal liver volume >1000 cc and no evidence of progressive or untreated extrahepatic disease. 46% of patients had received liver-directed therapy prior to SABR. Most lesions (82%) were treated with the full prescription dose of 75 Gy in 3 fractions, with the remainder receiving 52.5 to 67.5 Gy. At the time of SABR, 21 (34.4%) patients presented with stable extrahepatic disease and 40 (65.6%) had no evidence of other sites of metastases. The LC rates at 1, 3 and 5 years were 94%, 78% and 78%, respectively (Figure 2). Cancer type was not related to local relapse, although a subgroup of breast and gynaecological cancer patients had very favourable LC rates of 86%. Median survival was 27.6 months (Figure 3).
Figure 2: Local control
© Radiat Oncol 2018r
Figure 3: Overall survival
© Radiat Oncol 2018r