Hypofractionated three-dimensional conformal radiation therapy for primary liver carcinoma




The purpose of the current study was to evaluate the tolerance and efficacy of hypofractionated three-dimensional conformal radiotherapy (3DCRT) with or without transarterial chemoembolization (TACE) for technically unresectable or medically inoperable primary liver carcinoma (PLC).


Between April 1999 and August 2003, 128 patients with a clinical diagnosis of PLC received hypofractionated 3DCRT at Cancer Hospital, Guangxi Medical University. Both hypofractionated 3DCRT and TACE were used to treat 48 of these 128 patients. Liver cirrhosis of Child–Pugh Grade A was found in 108 patients, and Grade B was found in 20 patients. The mean gross tumor volume (GTV) was 459 ± 430 cm3. A mean total irradiation dose of 53.6 ± 6.6Gy was delivered at an average fraction of 4.88 ± 0.47Gy, 3 times a week using 8-MV photons.


The median follow-up time after 3DCRT was 12 months (range, 2–56 mos.). The immediate response rate was 55%. The overall survival rates at 1, 2, and 3 years were 65%, 43%, and 33%, respectively, with a median survival of 20 months (range, 7–31 mos.). Radiation Therapy Oncology Group (RTOG) Grade 2 acute gastrointestinal complications developed in 8 patients, whereas 4 patients developed Grade 3 late gastrointestinal complications. Radiation-induced liver disease (RILD) developed in 19 (15%) patients, of which 12 had Child–Pugh Grade B liver cirrhosis, and 7 had Grade A. GTV and associated liver cirrhosis were identified by Cox regression analysis as independent predictors for survival (P = 0.044 and 0.015).


Hypofractionated 3DCRT is effective in carefully selected patients with PLC. Gastrointestinal complications and RILD were the most distinct complications. Cancer 2005. © 2005 American Cancer Society.

Primary liver carcinoma (PLC) is one of the major cancers in the world with a mortality of more than 250,000 cases yearly. More than 137,000 cases of PLC were diagnosed yearly in China, which accounted for approximately > 40% of the total number in the world. PLC has become the second leading cause of death by cancer in China since 1990, and its annual mortality rate was 21.2/100,000 in the year 2000. Even though progress has been achieved for PLC diagnosis and treatment, its 5-year mortality rate is still > 95%.1, 2, 3 The management of technically unresectable and medically inoperable PLC remains challenging despite such therapies as transcatheter arterial chemoembolization (TACE), percutaneous ethanol injection therapy, microwave coagulation therapy, radiotherapy (RT), and liver transplantation. TACE showed improved patient survival, but its antitumor effect was frequently incomplete.4, 5 RT for PLC has resulted in unsatisfactory outcomes for the past three decades because poor tolerance of the liver to irradiation has prevented delivery of a tumoricidal doseage.6 Therefore, in literature, RT had no role in the management of PLC. Since the modern radiation technology of 3-dimensional conformal radiation therapy (3DCRT) and intensity-modulated radiation therapy (IMRT) was applied a decade ago, the role of RT in the management of PLC has been reconsidered. Investigations have demonstrated that partial liver irradiation alone or combined with nonsurgical modalities effectively induces tumor regression. However, massive normal liver tissue damage after external beam irradiation for large volume or diffuse PLC has been reported.7, 8, 9, 10 The technique of 3DCRT aided by a computerized treatment planning system has enabled tight conformation of a high dose volume to PLC lesions in 3-dimensions. Thus, this technique can spare normal liver tissue and is worth trying for PLC.11, 12 Several series that employed 3DCRT reported a dose–response relation in RT for liver cancers with better response rates and prolonged hepatic control in groups that received higher RT doses.9, 13 Hypofractionated RT, using large fraction size with an interval of > 1 day, has also been applied to liver tumors.14 Our hospital has, thus, incorporated 3DCRT with hypofractionation to treat PLC patients with the aim of delivering a high radiation dose to the tumor. We retrospectively analyzed the efficacy and adverse effects of this technique for 128 patients with PLC.



Patients were chosen from a database. Eligibility requirements included the following:

  • 1Patients with PLC had been diagnosed based on clinical diagnostic criteria proposed by the Chinese Society of Liver Cancer and the Chinese Anti-Cancer Association (CACA) in 1999.15 Briefly, clinical diagnosis of PLC can be made when patients meet all of the following criteria. 1) Level of α-fetoprotein (AFP) in serum must be elevated to > 400 μg/L, or 200 to < 400 μg/L with the percentage of Lens culinaris agglutinin reactive AFP < 25%. This raised AFP level should remain for more than 2 months. 2) Patients with non-PLC diseases or a condition that causes an increase in AFP should be excluded, including those patients with active hepatitis, pregnancy, or germ cell tumors, etc. 3) Patients associated with typical clinical symptoms of PLC. 4) Diagnosis of PLC had to be supported by evidence of radiologic imaging, i.e., computed axial tomography (CAT) and/or magnetic resonance imaging (MRI). 5) Metastatic liver cancers must be ruled out.
  • 2Patients' Eastern Cooperative Oncology Group (ECOG) performance status (PS) was 0–1.
  • 3Patients had received hypofractionated RT by 3DCRT technique.
  • 4All patient clinical data and follow-up data were available.

Between April 1999 and August 2003, 128 eligible patients were collected from patient files in the Department of Medical Records, Cancer Hospital, Guangxi Medical University, Guangxi Zhuang Autonomous Region, China. All patients' hospital charts and irradiation documents were carefully reviewed. Clinical characteristics of selected patients are described here. The group of 128 patients was comprised of 113 men and 15 women with a median age of 48.6 years (range, 23–77 yrs). Among these, there were 83 cases of T3 and 45 cases of T4 (UICC/AJCC, 1997). All 128 lesions were single. No patient had intrahepatic spreading. No patient had regional lymph node or extrahepatic metastases. Liver cirrhosis of Child–Pugh Grade A was documented in 108 patients, and 20 patients had Grade B. Portal vein thrombosis (PVT) on CAT was present in 34 patients and absent in 94 patients. Positive Hepatitis B virus (HBV) antigens appeared in 93 patients, and 35 patients had negative results for HBV. All patients had an ECOG PS of 0–1 (Table 1).

Table 1. Univariate Analysis of Prognostic Predictors for Survival in 128 PLC Patients Treated by 3DCRT with or without TACE
PredictorNo. casesOverall survival rate (%)χ2Log rank test P value
1 yr2 yr3 yr
  1. GTV: gross tumor volume; HBV: hepatitis B virus; TACE: transcatheter arterial chemoembolization. PVT: portal vein thrombosis; BED(Gy10); biologically effective dose calculated by L-Q model.

Age (yrs)      
 < 5592644435  
 ≥ 55366338200.670.413
T stage      
GTV (cm3)      
 < 125291008975  
 > 100012220037.310.000
Child-Pugh grade      
Fraction size (Gy)      
 < 565502525  
 > 6137355505.610.060
 < 7012472828  
 > 80516040340.010.940

During the study period, PLC patients who received 3DCRT generally had a PLC lesion that was technically unresectable, because of local advancement, or was medically inoperable because of severe liver cirrhosis. According to physician judgment, patients did not have other severe diseases that would interfere with RT. The PLC lesion margin was distinct on CAT and/or MRI. Informed consent had been obtained before treatment.


All 128 PLC patients received hypofractionated 3DCRT, and 48 out of 128 patients were treated by TACE before irradiation. The interval between TACE and 3DCRT was about 2 weeks.


TACE was performed with infusion of a mixture of 5 mL of lipiodol, 50–60 mg/m2 of epiadriamycin, and 30–40 mg/m2 of cis-platinum, or 5 mL of lipiodol, 6–7 mg/m2 of mitomycin C and 10–15 mg/m2 of 10-hydroxycamptothecin (HTCP), followed by Gelfoam® (Jinling Pharmacetical Co., Ltd., Nanjing, China), embolization. A median cycle of 2 (range, 1–4) was administered to 48 patients.


Patients were immobilized by a stereotactic body frame (SBF) and trained to breathe as shallowly as possible with the assistance of a tight abdominal belt to limit breathing. All patients underwent a planning CAT scan to facilitate three-dimensional treatment planning using the Topslane planning system (Topslane Medical Corp., Shanghai, China). The growth tumor volume (GTV) was delineated on CAT scan. Planning target volume (PTV) was determined by adding 0.5cm–1.5 cm to GTV. Liver motion was taken into account, and additional margins were added to account for motion. With the help of beam's eye view, four to eight coplanar or noncoplanar fields were designed. The dose was prescribed to isocenter as 100% without inhomogeneity tissue correction, and PTV was calculated to cover 90% of the isodose curve. Hypofractionated irradiation was applied, i.e., three fractions per week, i.e., on Monday, Wednesday, and Friday, respectively. As to the fraction size and total dose, these were determined on the basis of physicians' own judgments, which included patient PS and liver function. The fraction size ranged from 4 Gy to 8 Gy, and the total dose range was 40–60 Gy. However, the principle of “trial and error” had been observed during the study. At the very beginning, a large fraction, 7–8 Gy, was delivered, but more acute adverse effects and complications were noticed. Therefore, fraction size was decreased gradually to increase patient tolerance, but high total doses still remained. Late experience demonstrated that 4–5 Gy/fraction, 3 times a week to a total of about 50 Gy was tolerable. Thus, this fractionation was used for all patients in the later period of the study.

Before irradiation was implemented, portals were verified on simulator fluoroscopy to confirm the full coverage of a PLC lesion. 3DCRT was carried out by 8-MV photon linear accelerator (Philips, Best, the Netherlands).

For 128 patients, the average GTV was 459 cm3 ± 430 cm3. The total dose finally given was 53.6 Gy ± 6.6 Gy at a median fraction size of 4.88 Gy ± 0.47 Gy during median irradiation time of 4 weeks (range, 3–5 wks). Because of differences in fraction size and total dose, comparisons of radiation effect were difficult to determine between different fractionations. Therefore, to standardize the biologic effective dose (BED), the original linear quadratic model (L–Q) was used, which was as follows: BED = nd [1 + d/(α/β)]. Here, n was the number of fractions, and d was fraction size. A value of 10 Gy was chosen for α/β.

After standardization, < 70 Gy10 was given to 12 patients, 70–80 Gy10 to 65 patients, and > 80 Gy10 to 51 patients with a median BED of 79 Gy10 for the entire group, which was equivalent to a total dose of 52.7 Gy by conventional fractionation of 2 Gy/fraction and 5 fractions a week.

Adjuvant therapy

During treatment, combined adjuvant medications were given to patients to protect liver function.These medications consisted of vitamins, glutathione (GSH), and Chinese herbs.

Follow Up

Physical examinations, complete blood cell counts, and blood chemistry analyses were performed weekly during the 3DCRT alone or TACE and 3DCRT period. Patients were followed up every 3-months for 1 year, then every 6 months after completion of treatment. Each visit included history, physical examination, complete blood count, serum AFP, blood chemistry, abdominal ultrasound, and CAT scan.

Criteria for Toxicity

Irradiation toxicities were scored by RTOG criteria, including the acute, which occurred within the first 90 days of treatment, or the late, which occurred after 90 days of treatment. Acute liver toxicity was graded according to common toxicity criteria (CTC), National Cancer Institute (National Institutes of Health, Bethesda, MD). Late toxicity was focused on radiation-induced liver disease (RILD), which was proposed by Lawrence.16 RILD manifested by anicteric elevation of alkaline phosphatase (AKP) level of at least twofold and by nonmalignant ascites and hepatomegaly in the absence of documented progressive disease, or by elevated transaminases of at least fivefold the upper limit of normal or of the pretreatment level.

Endpoint for Outcome and Statistics

Immediate responses were classified according to World Health Organization (WHO) criteria by serial CT scans 3 months after 3DCRT. Complete disappearance of the tumor was scored as a complete response (CR); a decrease of > 50% in tumor size was scored as partial response (PR); a decrease of < 50% or an increase of < 25% in tumor size was scored as stable disease (SD); and an increase of > 25% or presence of a new lesion was scored as progressive disease (PD). Survival was estimated from the date of the beginning of treatment until death by the Kaplan–Meier model. Univariate and multivariate analyses by the Cox proportional regression model were performed to identify the prognostic factors important to survival.


Follow Up and Patients' Status at the Last Follow-Up Visit

During the period of treatment and at 3 months after treatment, the follow-up visits were well documented, but after completion of TACE and irradiation, follow ups were not consistent because periodical follow-up visits were not strictly observed because patients from remote rural areas had financial and transportation problems. However, status of survival or death was correctly recorded because all patients had been followed up through the last follow up, but concerning locally hepatic and distant failures, the time when these failures manifested was not accurately known. The last follow up was performed in December of 2003 with a median follow-up time of 12 months (range, 2–56 mos) for entire group. At the last follow-up visit, 63 patients were still alive, and 12 of these had no evidence of disease (NED); 51 of the living patients still had tumors, of whom 13 were scored SD; 24 patients had local tumor progression or intrahepatic spreading; 12 patients had accompanying distant metastases; 2 patients failed both locally and distantly. PLC caused the death of 46 patients, and, of these, 36 patients died from intrahepatic metastases, and 10 patients died from distant metastases. RILD caused the death of 16 patients, and 3 patients died from intercurrent diseases.

Tolerance and Toxicities

Acute treatment-related toxicities were assessable in all 128 patients. Nausea, fatigue, and loss of appetite occurred in 46 patients (36%) during the treatment period. Pain in the right upper abdomen (Grade 1–2) was noted in 13 patients (10%) at the beginning of RT. More acute toxicities were recorded in patients irradiated by a large fraction size of 7-8 Gy with morbidity of 77% (10/13 patients). However, these acute side effects could be relieved by a 20% dose of mannite, dexamethasone, and antalgica. Thrombocytopenia Grades 1–2 appeared in 18 (14%) patients, and Grade 3 appeared in 1 (1%) patient. Grades 1–2 leukopenia were seen in 3 (2%) patients. Grade 2 gastroenteritis occurred in 8 (6%) patients with gastrointestinal bleeding in occurring in 2 cases. Acute hepatic toxicity was noted in 47 (37%) patients with Grade 1 and in 14 (11%) patients with Grade 2, but all recovered eventually.

RILD was observed in 19 patients > 4 weeks after completion of irradiation with a median latent period of 5.5 weeks (range, 4–8 wks). Twelve RILD were Child–Pugh B patients with an incidence of 60% (12/20 patients), and 7 patients were Child–Pugh A with an incidence of 6% (7/108 patients).). After appropriate treatments, 16 of these 19 RILD patients died of hepatic failure unfortunately, and 3 of the 19 were still alive at the last follow-up visit. Univariate analyses demonstrated that T Stage, GTV, PVT, and Child–Pugh Grade of liver cirrhosis correlated with RILD (P = 0.024, 0.002, 0.001, and 0.000, respectively) in favor of T3 Stage, small GTV, no PVT, and Child–Pugh Grade A. Multivariate analysis found the Child–Pugh Grade of liver cirrhosis to be the sole independent factor predicting RILD (P = 0.000; RR = 14.8). For Child–Pugh Grade A PLCs, a mean dose of normal liver was critical with a tolerance dose of ≤ 19 Gy. When a mean radiation dose to the liver was ≤ 19 Gy, no RILD occurred, whereas RILD did occur in 6% of patients who were administered a dose > 19 Gy. Whereas, for patients with Child–Pugh Grade B, 60% (12/20 patients) manifested RILD. A separate analysis has been presented elsewhere.17

Another severe late complication was Grade 3 gastroduodenal ulcer in 4 (3%) patients, but no severe consequence resulted.

Tumor Response

Immediate tumor response was evaluated 3 months after 3DCRT for 121 patients of the original 128 (7 patients died within 3 months after treatment). Eight (7%) patients achieved CR, and 59 (49%) patients achieved PR with an objective response (CR + PR) rate of 55%. SD was observed in 17 (14%) patients, and PD was seen in 37 (31%) patients.

Overall Survival

The overall survival rates at 1, 2, and 3 years were 65%, 43%, and 33%, respectively (Fig. 1) with a median survival time of 20 (range,7–3l mos; 95% confidence interval [CI]) months.

Figure 1.

Graph depicts overall survival of 128 PLC patients treated by hypofractionated 3DCRT with or without TACE.

Prognostic Factors Affecting Overall Survival

Prognostic factors evaluated by univariate analysis consisted of gender, age, T Stage, GTV, HBV, TACE, PVT, Child–Pugh Grade, fraction size, and BED. Unluckily, gender, age, HBV, TACE, fraction size, and BED had no statistically significant impact on overall survival (all P > 0.05). However, T3 Stage, small GTV, no PVT, and Child–Pugh Grade A did predict a better overall survival (P = 0.001, 0.000, 0.000, and 0.0000, respectively) (Table 1 and Figs. 2–5).

Figure 2.

Graph depicts overall survivals and T Stages in 128 PLC patients treated by 3DCRT with or without TACE.

Figure 3.

Graph shows overall survivals and GTV (cm3) in 128 PLC patients treated by 3DCRT with or without TACE.

Figure 4.

Graph depicts overall survivals and portal vein thrombosis (PVT) in 128 PLC patients treated by 3DCRT with or without TACE.

Figure 5.

Overall survivals and Child–Pugh Grade in 128 PLC patients treated by 3DCRT with or without TACE are shown in this graph.

From Cox proportional regression analysis, GTV and liver cirrhosis were independent predictors for overall survival (P = 0.044 and 0.015, respectively) (Table 2). Patients with small GTV or Child–Pugh Grade A survived longer than others.

Table 2. Multivariate Analysis of Prognostic Parameters for Survival in 128 PLC Patients Treated by 3DCRT with or without TACE
ParameterβSEWaldSig.Exp(β)95% CI for Exp (β)
  1. CI: confidence interval; GTV: gross tumor volume; PVT: portal vein thrombosis.

T stage0.3500.4750.5430.4611.4190.559–3.598
Child-Pugh grade0.9280.3815.9330.0152.5281.199–5.333

Local Failure and Distant Metastases

At the last follow-up visit, hepatic failure, including local progression or intrahepatic metastases, was recorded in 62 (48%) patients. Distant metastases developed in 24 (19%) patients, of whom 17 had metastases in lungs, 3 in bones, and 4 in other sites.


PLC is one of the common cancers in China. Because there has not been a mass screening program for PLC in most provinces of mainland China, only 20% of PLCs can be treated by surgery when diagnosed. For the majority of PLCs, no effective treatment modality has been available, although TACE and tumor ablation by radiofrequency are still widely in use. Irradiation had had no place in the management of PLC, as demonstrated by the fact that there were no chapters on irradiation for PLC in most well known text books of radiation oncology or oncology. In the 1970s, we had tried radiation therapy for PLC but failed to show the predominant benefit because we had neither efficient means to identify gross tumor nor the equipment to perform proper irradiation techniques. Although we used a moving strip technique at that time to increase irradiation dose, the total dose delivered was approximately 30–40 Gy, which was far from the tumorcidal dose.18 Therefore, the attitude toward the use of RT for PLC had been negative until 3DCRT was introduced in recent years. Irradiation is now more accurate as 3DCRT can concentrate radiation on tumors, achieving a high conformity to tumor margins and sparing normal organs adjacent to tumors. In addition, a 3DCRT planning system has been facilitated by advancements in technology, including 3-dimensional reconstruction of tumors and of organs at risk, field design by beams-eye-view, and plan optimization, all of which have caused the role of RT in the management of PLC to be reconsidered. In the past decade, 3DCRT has been tried for PLC in a series of trials that have demonstrated an encouraging outcome. Matsuura et al. revealed a 2-year survival rate of 36.4% in 22 PLC patients treated with RT alone or in combination with TACE and percutaneous ethanol injection therapy.7 The Michigan group reported 36 cases of intrahepatic malignancies treated by RT with a median survival of 11 months.11 Recently Cheng treated primary hepatocellular carcinoma by RT, with or without TACE, and obtained a most encouraging result of a 41% 2-year survival rate and a median survival time of 19 months.4

Just five years ago, the technique of 3DCRT was established in our hospital. We began to try it for PLC patients who required it and whose physicians demanded it, but, at that time, we used 3DCRT only for the goal of palliation because of poor results from widely used TACE and other ablation techniques. These events have brought us to the present retrospective study.

First, it should be discussed that 128 PLC patients were diagnosed on clinical bases without histologic or cytologic evidence. Because of the fatal nature of PLC, especially for Stages T3–4, most PLC patients in Guangxi Zhuang Autonomous Region were not convinced to accept biopsy, which was invasive and associated with the risk of acute complications. Therefore, we used criteria for clinical diagnosis proposed by CACA. The experience in China has demonstrated that these criteria reached a diagnostic accuracy of > 95% compared with histology or cytology.15

In the current study, hypofractionated irradiation was used for all patients. The reason for this was to relieve patients' burden physically and economically. During this period, the determination of fraction size and total dose for individual patients was based on trial and error. At the very beginning of our study, a large fraction of 7–8 Gy was adopted for the first 13 patients, but as stated before, patient tolerance was poor with 77% (10/13 patients) experiencing nausea, fatigue, loss of appetite, and pain in the right upper abdomen. Therefore, fractionation size was decreased to 4–6 Gy, 3 times a week to a total radiation dose of approximately 50 Gy later in the study. Our experience later in the study demonstrated that hypofractionated irradiation could be well tolerated with fewer acute adverse effects; hence, it has been conventionally used for PLC since then.

All acute treatment-related toxicities, including gastrointestinal, hepatic, and hematopoietic were manageable without severe sequelae. However, acute as well as late complications in the gastrointestinal system should have been paid more attention when the 3DCRT plan was designed, especially for those tumors located in the left lobe of the liver. Nevertheless, the most serious late complication was RILD, which occurred 19 (15%) patients. Once RILD developed, it was almost always fatal with mortality of 84%. Univariate and multivariate analyses have demonstrated that liver cirrhosis was the most critical determinant for liver tolerance. As shown in the current study, patients with Child–Pugh Grade B cirrhosis were absolutely not able to tolerate the hypofractionated 3DCRT used in this study. However, for Child–Pugh Grade A patients, 3DCRT was tolerable if the mean dose to the liver was only < 19 Gy.

Our retrospective study proved that hypofractionated 3DCRT alone or combined with TACE for inoperable or unresectable PLC resulted in an improved 3-year survival rate of 33% and a longer median survival time of 20 months compared with historically reported rates. Of course, patients in this study were highly selected with good ECOG performance scores and without serious cirrhosis (84% patients with Child–Pugh Grade A). These encouraging results confirmed our postulation that 3DCRT could concentrate a radiation dose to a tumor with good conformity to the tumor margins and, therefore, spare normal liver tissue. The median dose to tumor for the entire group was equivalent to a total dose of 52.7 Gy by conventional fractionation, although a truly biologic dose would be less because the original L–Q model did not take tumor cell proliferation into account, which certainly lessened the biologic effect. Nevertheless, a dose of > 50 Gy was obvious higher than that used in the 1970s by conventional irradiation techniques. It was also believed that the radiosensitivity of PLCs was similar to that of epithelial carcinomas. Thus, more radiation should result in stronger inhibition of PLC, although it was not a tumoricidal dose.

As to prognostic predictors for PLC irradiation, tumor size, stage, serum AFP level, and several hepatic function-related serum parameters have been reported in literature.4, 9, 19 In our study, from univariate analysis, UICC/AJCC T Stage, GTV, PVT, and liver cirrhosis impacted significantly upon overall survival (P = 0.001, 0.000, 0.000, and 0.0000, respectively), in favor of smaller tumors, less GTV, no PVT, or Child–Pugh Grade A. Irradiation dose was also tested by univariate analysis for prognoses. To standardize irradiation effect, we applied the original L–Q model, which was obviously not reasonable because the proliferation of survived tumor cells during irradiation was not taken into consideration in this model. However, in literature, we were not able to find available parameters that described or explained proliferation kinetics of PLC during irradiation. Thus we used the L–Q model just for comparisons between different fractionations. Unluckily, dosimetric parameters did not provide any significant predictors. From Cox regression analyses, only GTV and liver cirrhosis were shown to be independent prognostic factors (P = 0.044 and 0.015).

Hepatic failures still accounted for a majority of deaths, with 48% due to local progression or intrahepatic spreading. Local control and intrahepatic metastasis rates could not be obtained in the current study by the Kaplan–Meier model because of shortcomings in follow up as mentioned above. However, it is noteworthy that the models we used show that treatment for local tumors and the liver as a whole should be more aggressive in the future. For distant metastases, it seemed unwise to use chemotherapy for PLC because patients generally associated this therapy with liver cirrhosis.

In the current study, in 48 patients, 3DCRT was combined with TACE, a therapy which has actually been widely used in China for unresectable PLC but not very efficiently for hepatic tumor control. Our purpose for the combination was based on the following considerations: 1) Combination of TACE and 3DCRT enhances tumor eradication effect; 2) PLC is well known for its biologic propensity toward intrahepatic metastasis or multicentral lesions in the liver. Thus, we relied on TACE to decrease intrahepatic spreading, or, at least, to delay its occurrence; 3). The margin of PLC lesions on CAT was not very sharp or clear in general. After TACE, in which iodine had been injected into tumor for embolization, the margins of lesions in the liver became very sharp. This made delineation of GTV easy and accurate. Unluckily, our study has failed to demonstrate the benefit of adding TACE to 3DCRT for increased survival rates compared with patients who were treated without TACE, but this study did show TACE's advantage in delineation of GTV. A criticism for administration of TACE before 3DCRT came up with the worry about increased production of hypoxic tumor cells after TACE because of blood vessel embolization. Theoretically this hypothesis was probably correct, but it was also clear that PLCs began to regrow shortly after TACE in most patients, which implied an improvement in blood supply to the tumor. Moreover, it is already known that one of the biologic characteristics of PLC is its association with a large percentage of hypoxic cells or, even, necrosis, especially for a large lesion. Hence, further investigation is necessary to determine whether TACE worsens hypoxic conditions and to what extent this may occur.

For future clinical trials of PLC treatment by 3DCRT, we propose the following: 1) Indication: 3DCRT could be recommended for medically inoperable or technically unresectable PLCs, especially in patients with a small tumor and no serious liver cirrhosis (Child–Pugh Grade A). 2) Fractionation and total dose: Liver is thought of as a late-responding normal organ, so that large fraction size may increase hepatic irradiation toxicity. Because of the unexpected good result in the current study, it seems wiser to use a conventional fractionation of 2 Gy per fraction, 5 fractions a week, to a total accumulated dose of 50 Gy. However, a dose escalation study should be carried out to determine the maximum tolerance dose. 3) The use of TACE is suggested before irradiation for combined therapeutic effect. 4) The 3DCRT technique needs to be improved. First is the delineation of GTV. In addition to TACE, fusing images produced by MRI and CAT scans is another excellent technique, according to our preliminary experience, for tumor delineation. Second is the motion of a PLC lesion as a patient breathes. We have recently begun to use an active breath coordinator (ABC) to obtain greater accuracy.

In conclusion, treatment with hypofractionated 3DCRT resulted in substantial tumor regression and prolongation of survival in carefully selected patients who had an unresectable or inoperable PLC provided that greater awareness of gastrointestinal complications and RILD was taken into consideration.


Authors would like to thank Dr. Jin-Hui Liu for her assistant in editing text in English.