Efficacy and safety of transarterial chemoembolization combined to conformal radiotherapy for uninodular hepatocellular carcinoma

Purpose: A proportion of patients with uninodular hepatocellular carcinoma (HCC) cannot benefit from potential curative therapies such as liver transplantation, surgical resection or radiofrequency ablation. Thus, they are prone to receive transarterial chemoembolization (TACE) that is a palliative option with low probability of both complete response and prolonged local control. Herein, we assessed the combination of TACE and 3D-high dose conformal radiotherapy (3D-HDCRT) for efficacy and safety in HCC. Methods: We retrospectively analyzed the outcome of 35 consecutive patients with uninodular HCC ≤ 100 mm, treated by one course of TACE combined to 3D-HDCRT. The follow-up consisted on clinics, biology, hepatic CT-scan or MRI at month-1 and -3, and thereafter every 3 months. Results: Complete response was obtained in 80% of patients following mRECIST criteria (95% in HCC ≤ 50 mm, and 60% in HCC > 50 mm) with uncommon local recurrence (11%), overall survival rates of 79%, 59% and 44% at respectively 1, 2 and 3 years (median, 37.3 months), and 11.4% grade-3/4 toxicities. Pre-therapeutic α-fetoprotein level ≥ 200 ng/mL was found as a strong predictor of poorer outcome. Conclusion: We showed that TACE combined to 3D-HDCRT can be highly efficient to reach local control and interesting overall survival rates for uninodular HCC, with limited severe toxicities for Child-Pugh A patients. Subsequent prospective controlled trials are warranted for comparison with therapeutic standards.


Introduction
Hepatocellular carcinoma (HCC), one of the most common cancer worldwide, is the main cause of mortality among cirrhotic patients, and ranks third in terms of death by cancer. HCC prognosis greatly varies according to tumour size and spread as well as liver functions and general status. 1,2 Except a restrictive population of very early or early stage HCCs that are eligible for curative options such as orthotopic liver transplantation (OLT) or radiofrequency ablation (RFA), the best prognosis remains with surgical resection. 3 This latter can be considered as a curative option, especially for patients with uninodular HCC, and its feasibility is closely linked to the tumour bulk, absence of clinically relevant portal hypertension, and conserved liver functions. Tumour size is not a limiting issue per se, but the best outcome post-surgery remains for patients with HCC ≤ 5 cm. However a pattern of concerns arises from uninodular HCCs since they might not be eligible for curative options, and thus are candidate for trans arterial intrahepatic chemoembolization (TACE).
TACE is a palliative option with the best reported median overall survival around 20 months (mo). TACE induces extensive tumor necrosis in more than 50% of HCCs, but most of tumors do not achieve a complete response (around 20% complete responses following liver-adapted RECIST criteria), and TACE is of limited efficacy in terms of progression-free survival. Indeed, within or around the capsule which is supplied by both arterial and portal blood, tumour cells remain viable and are frequently responsible for subsequent progression, this consideration prompting treatment repetitions. As the tumour response variable appears as an independent predictive factor of survival, it seems of likelihood that com- 2 Merle et  plementary approaches tending to increase complete necrosis could be of benefit on survival. [4][5][6][7][8][9][10] Combination of three-dimensional high-dose conformal radiotherapy (3D-HDCRT) and TACE may remedy the limitation of TACE alone by adding anti-cancerous synergy, and serving the purpose of residual cancer cell eradication after TACE. All the so far published studies on TACE + 3D-HDCRT combination are uncontrolled trials or retrospective analysis. Initial studies tested norm fractionated schedules of 3D-HDCRT following TACE (25-69.3 Gy, 1.6-2 Gy/daily fraction), leading to interesting response rates (49 to 63%), and around 18 months overall survival. Of interest, toxicities were reduced, including only transient elevation of liver function tests in all patients, but without treatment-related death. 11,12 Subsequent studies compared TACE + 3D-HDCRT combination to TACE alone in retrospective analysis, and clearly isolated irradiation as an independent predictor of tumour response and prolonged survival. [13][14][15] In the present study, we showed that, for uninodular HCC developed in noncirrhotics or Child-Pugh A cirrhotics, the combination of TACE and 3D-HDCRT led to high complete response rates, low risk of local recurrence and acceptable toxicity.

Study Characteristics
In our centre, every HCC patient is mandatory reviewed by a weekly multidisciplinary board composed of liver surgeons and transplanters, interventional radiologists, medical oncologists, radiotherapists and hepatologists. Therapeutic strategies are defined following international guidelines. 7 We retrospectively analyzed the outcome of 35 consecutive patients treated in our centre by TACE + 3D-HDCRT combination. All the patients fullfiled the following criteria: histologic or cytologic proven and/or noninvasive imaging-based diagnosed HCC using AASLD criteria with CT-scan and/or magnetic resonance imaging (MRI) 7 ; solitary HCC nodule ≤ 100 mm developed within a nontransplanted liver; patient ≥ 18 years not suitable for standard curative options (OLT, surgical resection or RFA) following international guidelines 7 ; Child-Pugh A5 to B7 status if cirrhosis; absence of truncular or lobar portal vein invasion, or suprahepatic vein invasion; absence of previous liver irradiation; performance status < 2; all the patients received informations on the treatment schedule based on previously published data and freely consented for this therapeutic association.

TACE Procedure
One single course of TACE was performed. All the patients underwent baseline angiography of the celiac trunk, superior mesenteric artery and hepatic artery using a peripheral arterial approach. TACE was performed with suspension of Doxorubicin (50 mg) and 10 ml of lipiodol, before emboliza-tion with mixed absorbable gelatin sponge (Curaspon). After selecting the tumor-bearing arteries, and when possible for the most hyperselective treatment, a microcatheter was used for the drug infusion in the segmental or supra segmental arteries. Patients were discharged after recovery followed by physical examination and blood test 3 days after TACE. We contraindicated patients to arterial procedure when presence of impaired clotting tests (platelet count below 50 × 10 9 /L or prothrombin activity below 50%), renal failure, severe atheromatosis; any contraindication to doxorubicin (serum bilirubin > 50 µmol/L, leucocyte count < 3 × 10 9 /L, cardiac ejection fraction < 50%); or endstage tumoral disease -BCLC-D).

Radiation Therapy
3D-HDCRT radiation therapy started 10 days post-TACE 16 , including active breathing control ("respiratory gating"). 17 The treatment planning computed tomography (CT) scans were performed in the position of 3D-HDCRT. The lipiodol tumor uptake was an effective target to define the contours of the lesions. Nontumor and tumor volumes were outlined by an experienced radiation oncologist from data of inspiratory hold sequences as previously described. 18 As the gating technique was used, radiation was delivered during the inspiration phase. As previously reported, radiation doses were determined by the dose-volume histograms (DVH) 19,20 and normal tissue complication probabilities (NTCP). 21 Total dose was 54 Gy. However lower doses could be used -i.e. 45 or 30 Gy -depending on whether the volume of normal liver parenchyma receiving 50% of the isocenter dose was respectively comprised between 33% and 66%, or > 66%. Dose constraints were included on the stomach and duodenum. Fractionation was 3 Gy/fraction, 5 fractions/week, with photon beam energy > 10 MeV. Clinical, biologic and radiologic follow-up were performed 1 month after the end of radiotherapy, and subsequently every 3 months. The same dosimetric schedule was equally applied for both cirrhotic and noncirrhotic patients, taking account that all cirrhotic patients had rather well conserved liver functions (Child-Pugh A5 to B7). We did not use non coplanar fields routinely. Either 2 opposed fields were used, or a technique of 5 to 8 fields was used, according to the tumor location and tumor size. The proximity of stomach and duodenum was an important decision tool in the choice of the fields arrangement.

Assessment of Efficacy and Toxicity
Efficacy of the TACE + 3D-HDCRT combination therapy was assessed on: tumor response by injected dynamic CT-scan (or MRI) following liver adapted-RECIST criteria; overall survival (OS); progression-free survival (PFS); time to relapse (TTR) for complete responders (CR), or time to progression (TTP) for partial responders (PR) or stable diseases (SD); safety using both the Child-Pugh score for liver toxicity, the national cancer institute common toxicity criteria (NCI-CTC)

Statistical Methods
Probabilities of survival were calculated from the day of screening CT-scan (MRI) imaging according to Kaplan and Meier. Relationship between each of the variables and survival was assessed by log-rank test in univariate analysis. Relationships between the different variables were analyzed by the CHI-2 test (P value < .05 indicating significance). Multivariate analyses were performed using Cox models for overall survival, PFS and TTR/TTP. All statistical tests were two-sided and p-values < .05 were considered significant. Statistical analysis was performed using MedCalc version 12.3.0.0 software package.

Discussion
In the present study, we aimed to catch the antitumor effect and safety of TACE + 3D-HDCRT combination therapy. We focused on a specific subset of patients -i.e. those with uninodular HCC tumor ≤ 100 mm, ineligible for curative ther apies (OLT, surgical resection or RFA), and developed in chronic liver disease (mainly cirrhosis) devoid of significant hepatic insufficiency. We demonstrated that TACE + 3D-HDCRT can reach high rate of tumor response, interesting survival rates and fairly good tolerance. Therapeutic strategies are defined following international guidelines. 7 Patients non eligible for surgery are patients with: i) non-stage A HCC in the BCLC classification; ii) significative portal hypertension with esophageal varices and/or porto-sushepatic gradient > 10 mm Hg; iii) comorbidities forbidding general anesthesia for surgery. Patient non eligible for monopolar RFA were either patients with HCC tumor size > 30 mm diameter, or patients with < 30 mm diameter HCC with localization prohibiting percutaneous RFA (not visible at US examination, localized in the upper the part of the liver, in segment I, exophytic/subapsular, etc).
We chose to assess efficacy of TACE + 3D-HDCRT in uninodular HCCs ineligible for potential curative options. 3D-HDCRT is particularly relevant for HCCs which topography allows radiation beams at focalizing onto a small part of the liver and sparing the surrounding non-tumorous area. 23,24 In this way, it clearly appears that uninodular HCCs are the best candidates. In contrast, multifocal HCCs scattered in both hepatic lobes do not seem suitable due to the potential huge volume of irradiated non-tumorous liver parenchyma, thus prohibiting delivering high dose of radiation by risk of iatrogenic hepatitis. Herein, one single course of TACE was performed followed 10 days later by 3D-HDCRT to take advantage of the potential synergy between ionizing radiations and the cytotoxic chemotherapeutic agent (doxorubicin) inside the tumor. TACE induces hypoxic necrosis, followed by VEGF peak and regrowth of residual cancerous foci, this event happening early post-TACE. 25 Thus it seemed important to deliver radiation therapy before re-growth of the residual cancer cells. To this aim, we used high dosage of photon therapy (54 Gy, 3 Gy per fraction, equivalent to 65-70 Gy total dose with 2 Gy per fraction) as compared to those previously published: 69 Gy with 1.8 Gy per fraction, or 62 Gy with 2 Gy per fraction. 11,26-28    We showed that TACE + 3D-HDCRT combination therapy is highly efficient for tumor destruction. Tumor size  50 mm greatly predicted CR (95%) vs. 60% for HCCs > 50 mm, with uncommon relapse (11%) within radiation fields for CR pts. We previously reported that 3D-HDCRT monotherapy could lead to 80% CR for HCCs  50 mm, but local recurrence rates were higher (20%) than in the present TACE + 3D-HDCRT combination (11%). 18 Regarding this subset of HCCs  50 mm, we further showed that TACE + 3D-HDRT leads to one and 3-year overall survival rates of 80% and 54% (median, 39.6 months), quite similar to those reported in literature with percutaneous destructions or surgical resections. 29 Surgical resection was reported as treatment of choice for HCCs ≤ 50 mm, whereas OLT or RFA, these latter being more prone for smaller HCCs (≤ 30 mm), could be alternative options in curative intent to treat. In cohorts of subset of HCCs ≤ 30 mm, surgery is quite equivalent to RFA as illustrated by some studies where the one and 3-year survival rates were reported at 89% and 57% in surgical groups, and 91% and 71% in RFA groups (P = .30). However in the same study, surgery was more efficient than RFA for the subset of HCCs ranging from 31 to 50 mm: one and 3-year survival rates were reported at 92% and 68% in the surgical group vs. 81% and 42% in the RFA group (P = .03). [30][31][32][33][34][35][36] Unfortunately, significant number of patients with uninodular small size HCC cannot be treated by these potential curative options due to patient characteristics or tumour topography. Furthermore only a limited number of those patients can benefit of liver transplantation. They are thus intended to TACE that is a palliative option, giving less than 20% complete responses, and 20 months overall survival. 5 Regarding HCCs > 50 mm in the present study, TACE + 3D-HDCRT led to 64% CR, 79% and 42% overall survival rates at one and 3-years (median, 30 months). Some published series reported that surgical resection of HCCs > 50 mm gives 69% and 37% overall survival rates at respectively 1 and 3-years (median, 20 months) 37,38 , thus at least comparable, although with a trends to inferiority, to our TACE + 3D-HDCRT data.
In previously published studies, predictive factors of prolonged survival post-TACE + 3D-HDCRT were liver functions as assessed by the Child-Pugh score, tumor size (< 50 mm vs. 50-100 mm vs. > 100 mm), absence of neoplastic portal vein invasion, total dose irradiation delivered to the tumor (60 Gy vs. 56 Gy vs. 48 Gy with 1.8 Gy/fraction). 39 In the present study, almost all patients were Child-Pugh A, devoid of portal vein invasion and the irradiation dose was constant (54 Gy, 3 Gy/fraction). We found pre-therapeutic AFP < 200 ng/mL as predictive factor of prolonged survival, and at a lesser extent since nonsignificant, tumor size ( 50 mm) and CR. Furthermore, pre-therapeutic AFP ≥ 200 ng/mL and tumor size > 50 mm clearly appeared as predictive of tumor recurrence such as intra-hepatic metastasis since occurring mainly outside the radiation fields. Thus it's highly possible that pre-therapeutic AFP and tumor size predict the presence of micro metastasis spared in the liver parenchyma before TACE + 3D-HDCRT procedure. Although tumor response is likely a key factor for prolonged survival, this parameter did not appear as determinant in the present study since: i) most of patients were CR and only a few were PR or SD, thus prohibiting robust statistical comparisons; ii) all PR pts could subsequently benefit from additional RFA or TACE to obtain complete response on the residual living tumor tissue within the previously irradiated nodule.
Comparing our data with those previously published in the literature, the first study combining 3D-HDCRT (total 44-69.3 Gy, daily 1.8 Gy/fraction) and TACE was published in 1999 and enrolled patients with unresectable large size HCCs (mean 9 ± 3.4 cm). The objective response rate was 63.3%, and survival reached 67% and 22.2% at one and 3-years, with 17 months overall survival. 11 Although achieved in uncontrolled settings, subsequent studies compared TACE+3D-HDCRT to TACE monotherapy, and clearly isolated irradiation as an independent positive predictor of survival. The combination therapy gave higher response rates than TACE monotherapy (47.4% vs. 28.1%, P < .05), and additionally prolonged survival (19 vs. 10 months, P = .0001). 13 Further studies evaluated increasing doses of radiation therapy on HCC in chemo-radiation strategies (up to 60 Gy in 7.5 Gy/fraction, equivalent to 87.5 Gy in 2 Gy/fraction) delivered by conformational hypofractionation. This radiation schedule led to higher response rates (90.5%) and better overall survival reaching 25 months. 39 Regarding safety, our data are rather encouraging since showing only 11.4% grade 3-4 toxicities, whereas previously published studies reported huge variability in values: 13-50% grade-3 toxicities. 40 The limited high grade toxicity observed in our study may be explained by the strict selection of patients at inclusion (Child-Pugh A) with uninodular HCC (limiting widespread irradiation of the liver parenchyma) and adjacent organs such as stomach for instance.
In parallel of 3D-HDCRT, other approaches such as stereotactic body radiotherapy (SBRT) reported exciting data for inoperable HCCs. SBRT allows the delivery of higher doses of radiotherapy as compared to 3D-HDCRT, with more accuracy to target the tumor volume and to spare the nontumorous surrounding liver parenchyma. In Jang's et al report 41 , a 3 fraction 51 Gy SBRT (range, 33-60 Gy) led to 87% local control in a dose efficient manner for tumors ranging from 10 to 70 mm, mean 31 mm, median 30 mm (100% in our study, with tumors ranging from 23 to 88 mm, mean 49 mm, median 50 mm). The one, two and three year overall survival rates reached 82%, 63% and 55% respectively, median around 42 months with SBRT vs. 79%, 59% and 44%, median 39.6 months in our TACE+3D-HDCRT study. Grade 3/4 toxicities were quite similar between Jang's reports and the present study (10% vs. 11.4%). Prospective randomized controlled trials could compare each approach between SBRT and

Conclusion
Finally in the present study, it seems of interest to be aware that patients with HCC of poor prognosis due to ineligibility to surgical resection or RFA, might keep potential comparable efficiency for tumor control and survival probability thanks to TACE+3D-HDCRT combination therapy. However, randomized controlled studies are needed to definitely clarify this concern. Currently, the TACERTE study is under investigation. This is a multicentre, prospective, randomized, controlled phase 2B trial comparing TACE+3D-HDCRT vs.
TACE alone for HCCs < 10 cm and ineligible to curative options.