Radiation Induced Liver Toxicity

Abstract

Liver was always considered to be ‘highly sensitive’ to radiation therapy (RT) and was not considered ‘safe’ for radiation therapy treatment. The most significant radiation induced liver toxicity was described by Ingold et al. as “Radiation hepatitis.” Historically, radiation to liver lesions with curative intent or incidental exposure during adjacent organ treatment or total body irradiation implied whole organ irradiation due to lack of high precision technology. Whole organ irradiation led to classic clinical picture termed as “Radiation Induced Liver Disease (RILD).” In conventional fractionation, the whole liver could be treated only to the doses of 30–35Gy safely, which mostly serves as palliation rather than cure. With the advent of technological advancements like IMRT, especially stereotactic radiation therapy (SBRT), the notion of highly precise and accurate treatment has been made practically possible. The toxicity profile for this kind of focused radiation was certainly different from that of whole organ irradiation. There have been attempts made to characterize the effects caused by the high precision radiation. Thus, the QUANTEC liver paper distinguished RILD to ‘classic’ and ‘non-classic’ types. Classic RILD is defined as ‘anicteric hepatomegaly and ascites’, and also can also have elevated alkaline phosphatase (more than twice the upper limit of normal or baseline value). This is the type of clinical picture encountered following irradiation of whole or greater part of the organ. Non-classic RILD is defined by elevated liver transaminases more than five times the upper limit of normal or a decline in liver function (measured by a worsening of Child-Pugh score by 2 or more), in the absence of classic RILD. In patients with baseline values more than five times the upper limit of normal, CTCAE Grade 4 levels are within 3 months after completion of RT. This is the type of RILD that is encountered typically after high-dose radiation to a smaller part of liver. It is commonly associated with infective etiology. Emami et al. reported the liver tolerance doses or TD 5/5 (5% complication rate in 5 years) as 50 Gy for one-third (33%) of the liver, 35 Gy for two-thirds (67%) of the liver, and 30 Gy for the whole liver (100%). Liver function (Child Pugh Score), infective etiology, performance status and co-morbidities influence the radiation induced toxicity. Lyman–Kutcher–Burman (LKB)-NTCP model was used to assess dose-volume risk of RILD. Lausch et al. at London Regional Cancer Program (LRCP), developed a logistic TCP model. Quantitative Analysis of Normal Tissue Effects in the Clinic (QUANTEC) reported recommendations that mean normal liver dose should be <18 Gy for baseline CP-A patients and < 6 Gy for those with CP-B, for a 6-fraction SBRT regimen. The University of Colorado phase 1 clinical trial of SBRT for liver metastases described the importance of the liver volume spared, that is, ‘critical volume model.’ It is estimated that a typical normal liver volume is approximately 2000 mL and specified that a minimum volume of 700 mL or 35% of normal liver should remain uninjured by SBRT i.e. at least 700 mL of normal liver (entire liver minus cumulative GTV) had to receive at total dose less than 15 Gy. In treatment regimen of 48 Gy in 3 fractions, CP-A patients were required to either limit the dose to 33% of the uninvolved liver (D33%) < 10 Gy and maintain the liver volume receiving <7 Gy to <500 cc. In more conservative treatment regimen, such as in 40 Gy in 5 fractions schedule, CP-B7 patients had to meet constraints of D33% < 18 Gy and/or > 500 cc receiving <12 Gy. The concept of body surface area (BSA) and Basal Metabolic Index (BMI) guided estimation of optimal liver volume is required to estimate the liver volume need to be spared during SBRT treatment. Radiation induced liver injury is potentially hazardous complication. There is no definitive treatment and a proportion of patient may land up in gross decompensation. Usually supportive care, diuretics, albumin supplement, and vitamin K replacement may be useful. Better case selection will avert incidence of RILD. Precise imaging, contouring, planning and respecting normal tissue constraints are critical. Radiation delivery with motion management and image guidance will allow delivery of higher dose and spare normal liver and hence will improve response to treatment and reduce RILD.

1. Introduction

Liver was always considered to be ‘highly sensitive’ to radiation therapy and was not considered ‘safe’ for radiation therapy treatment. For many years, maximal tolerable dose (mean liver dose) for liver was considered to be low and radiation dose required for therapeutic effect for liver tumors (Hepatocellular carcinoma, Cholangiocarcinoma) was considered high. Hence, radiation therapy was mostly not considered for liver tumors. Liver is a moving organ and movement of liver is dependent on many factors such as breathing pattern, stomach filling, peristaltic movements, and hence liver movement is not predictable. There were no appropriate tracking technology or high dose radiation delivery technology in ‘moving’ targets like liver. Hence, in early years of radiation therapy there are only a few anecdotal reports of radiation therapy delivery in liver tumors. In recent years, with advent of motion management system and technology to deliver high dose of radiation therapy that has increased the usage and literature about radiation therapy in liver tumors regarding both response to treatment and toxicities. In liver tumors, radiation therapy is mostly recommended in hepatocellular carcinoma (HCC), intra-hepatic cholangiocarcinoma and liver metastasis.

The most significant radiation induced liver toxicity was described by Ingold et al. as “Radiation hepatitis” [1]. Historically, radiation to liver lesions with curative intent or incidental exposure during adjacent organ treatment or total body irradiation implied whole organ irradiation due to lack of high precision technology. This kind of whole organ irradiation led to a classic clinical picture which was then termed as “RILD.” In 1966, Reed et al. have worked on pathology of radiation injury to liver and have established that the early changes are obliteration of small vasculature followed by secondary effects such as hyperemia and cell loss. They have also concluded that there is effective re-establishment of hepatic vasculature and return of normal hepatic structure with time [2]. Liver consists of hepatocytes connected as parallel structures and hence liver is considered a ‘parallel’ structure. This means, even if a small portion of liver is damaged, other part of the liver will work as ‘parallel’ structure and there will be no functional damage. If a large portion of liver is damaged and hence the ‘parallel’ architecture is affected then there will be disruption of function. This means, mean dose to liver is critical, whereas maximum dose or small high dose to liver may not have clinical relevance. In conventional fractionation, the whole liver could be treated only to the doses of 30–35Gy safely which serves only the purpose of palliation rather than cure. This aspect had set radiation aside of the curative liver therapy for many decades. With the advent of technological advancements like IMRT, especially SBRT, the notion of highly precise and accurate treatment has been made practically possible. This enabled focusing high doses of radiation to the tumor, sparing the normal liver thus bringing back the option of radiation for liver lesions into light once again. With the use of these, a significant portion of liver could be saved from high doses of radiation. The toxicity profile for this kind of focused radiation was certainly different from that of whole organ irradiation. There have been attempts made to characterize the effects caused by the high precision radiation. Thus the QUANTEC liver paper distinguished RILD to ‘classic’ and ‘non-classic’ types [3].

1.1 Classic RILD

Defined as ‘anicteric hepatomegaly and ascites’, also can also have elevated alkaline phosphatase (more than twice the upper limit of normal or baseline value).

This is the type of clinical picture encountered following irradiation of whole or greater part of the organ. As explained by Reed and Cox [2], this is related to the hepatic vascular changes leading to hepatic parenchymal necrosis. Although it takes time, this phenomenon is seen to be reversible in most of the cases. But the repair of hepatic structure is dependent on the baseline liver status. The incidence described in the 2000s was 5–10% if mean liver dose constraint of 30–35 Gy was met.

1.2 Non-classic RILD

Defined by elevated liver transaminases more than five times the upper limit of normal or a decline in liver function (measured by a worsening of Child-Pugh score by 2 or more), in the absence of classic RILD. In patients with baseline values more than five times the upper limit of normal, CTCAE Grade 4 levels within 3 months after completion of RT.

This is the type of RILD that is encountered typically after high dose radiation to a smaller part of liver. It is commonly associated with infective etiology. Although the exact pathogenesis is un-clear, it involves loss of regenerating hepatocytes. This is usually not irreversible.

The characteristics of them are summarized in Table 1.

	Classic RILD	Non-classic RILD
Time to presentation post Rx	2 weeks to 3 months	1 week to 3 months
Prone candidates	Otherwise fairly well-functioning pre-treatment liver	Common in those with poor liver function (hepatitis B infection, Child-Pugh Classes B and C)
Patho-physiology	There is occlusion and obliteration of the central veins of the hepatic lobules, retrograde congestion, and secondary hepatocyte necrosis	Un-clear but involves loss of regenerating hepatocytes and reactivation of hepatitis
Jaundice	−	++
Ascites	+++	+
Laboratory findings
Increased Bilirubin	+	+++
Increased AST	2 times ULN	5 times ULN
Increased ALP	+++	+

Table 1.

Differences between classic and non-classic RILD.

2. Grade of toxicity

Common Terminology Criteria for Adverse Events (CTCAE) version 5 has graded various symptoms like hepatic pain, hepatic necrosis, hepatic hemorrhage distinctly. Hepatic failure is defined as a disorder characterized by the inability of the liver to metabolize chemicals in the body. Asterixis, mild encephalopathy is grade 3 whereas moderate to severe encephalopathy and coma is grade 4 and death is grade 5 (Table 2).

CTCAE term	Grade 1	Grade 2	Grade 3	Grade 4
Hepatic pain (Sensation of marked discomfort in the liver region)	Mild pain	Moderate pain; limiting instrumental ADL	Severe pain; limiting self-care ADL	—
Hepatic hemorrhage (Bleeding from liver)	Mild symptoms; intervention not indicated	Moderate symptoms; intervention indicated	Transfusion indicated; invasive intervention indicated; hospitalization	Life-threatening consequences; urgent intervention indicated
Hepatic necrosis (A disorder characterized by a necrotic process occurring in the hepatic parenchyma)	—	—	—	Life-threatening consequences; urgent invasive intervention indicated
Hepatic failure (A disorder characterized by the inability of the liver to metabolize chemicals in the body)	—	—	Asterixis, mild encephalopathy; drug induced liver injury; limiting ADL	Life-threatening consequences; moderate to severe encephalopathy; coma
Sinusoidal obstruction syndrome (A disorder characterized by severe hepatic injury as a result of the blood vessels of the liver becoming inflamed and/or blocked)	—	Blood bilirubin 2–5 mg/dL; minor interventions required (i.e., blood product, diuretic, oxygen)	Blood bilirubin >5 mg/dL; coagulation modifier indicated (e.g., defibrotide); reversal of flow on ultrasound	Life-threatening consequences (e.g., ventilatory support, dialysis, plasmapheresis, peritoneal drainage)

Table 2.

CTCAE Grading of liver toxicity.

Taking into account the wider applicability in cancer treatment, the CTCAE toxicity grading is non-specific to radiation induced toxicity. It does not take into consideration the performance status of the patient, baseline liver function and the relative changes in liver function caused by radiation, which is more clinically relevant and a predictor of reversibility of the RILD.

The Radiation Therapy Oncology Group (RTOG) liver toxicity grading includes nausea, dyspepsia as grade 1, abnormal liver function tests with normal serum albumin as grade 2, disabling hepatic insufficiency with low albumin, edema, ascites as grade 3 and necrosis, encephalopathy, hepatic coma as grade 4, death as grade 5 (Table 3) [4].

	Grade 0	Grade 1	Grade 2	Grade 3	Grade 4	Grade 5
Liver toxicity	None	Mild lassitude; nausea, dyspepsia; slightly abnormal liver function	Moderate symptoms; some abnormal liver function tests; serum albumin normal	Disabling hepatitic insufficiency; liver function tests grossly abnormal; low albumin; edema or ascites	Necrosis/Hepatic coma or encephalopathy	Death directly related to radiation induced late effects

Table 3.

RTOG/EORTC late morbidity grading.

The drawback of this grading system is again the lack of specificity in scoring of liver function tests.

3. Time of RILD and presenting symptoms

RILD can be an acute or sub-acute phenomenon. It typically occurs 4–8 weeks, but can occur 7–90 days post radiation [5]. Rarely, it is seen to occur as late as 7 months. Though there has not been much variation in time to presentation between classic and non-classic RILD, classic tends to occur earlier. The clinical manifestations of RILD are non-specific but patients typically present with symptoms like fatigue, weight gain, increased abdominal girth, rarely abdominal pain. There can be signs of hepatomegaly, ascites, altered liver function, elevated alkaline phosphatase disproportionate to other liver enzymes. In case of non-classic RILD, there can be jaundice and marked elevation of liver enzymes. RILD is essentially a diagnosis of exclusion. Radiologic sequel is seen as sharply demarcated low attenuation areas on CT. In case of steatotactic liver background, there can be areas of elevated attenuation. MRI can show areas of increased T2 signal keeping in with acute inflammation [6].

4. Evaluation parameters of radiation toxicity

Various empiric end points have been used to describe RILD, which include deterioration in Child Pugh score and RTOG/CTCAE grade 2–4 abnormal laboratory values. One such end point evaluated is Child Pugh (CP) score declining by 2 or more scores. Chapman et al. tried to define clinically relevant endpoints in cirrhotic patients post SBRT or proton beam therapy. In the retrospective review of 48 patients, multivariate analysis showed that Child Pugh Score increase of ≥1 or ≥ 2, CTCAE AST toxicity grade change were the strongest predictors of OS and RILD specific survival also [7]. This has been confirmed by other studies also. In a prospective study evaluating Child Pugh score as a tool for assessment of acute toxicity of liver SBRT, 94 patients were analyzed and 15% had RILD. In CP score assessment at 2 month follow up, 46 (38%) had no change in CP score. Decline of 1-, 2- & 3-point CP score from baseline was in 17%, 10%, 14%. Improvement in CP score of 1- & 2- point from baseline was in 9% and 1% respectively. CP score change after SBRT correlated with the post RT acute toxicities in the study and hence CP score change was considered as an objective scoring system to evaluate the radiation induced liver injury after SBRT treatment [8].

Other parameters include Model for End-stage Liver Disease (MELD) score, CLIP score, GRETCH score, albumin-bilirubin (ALBI) score, PIVKA, AFP grade (Figure 1).

Figure 1.

Time frame of different scoring system.

4.1 MELD scoring system

The MELD score is a chronic liver disease severity scoring system that is calculated from serum bilirubin, creatinine and INR, but modified to include serum sodium concentration (MELD-Na) [9]. It was originally developed to predict three-month mortality following transjugular intrahepatic portosystemic shunt (TIPS) placement. It is frequently used for patients being evaluated for transplant.

4.2 CLIP scoring system

The CLIP score includes Child-Pugh stage, tumor morphology and extension, serum alfa-fetoprotein (AFP) levels, and portal vein thrombosis [10]. It takes into account both liver function and tumor characteristics and has been validated for HCC staging in relation to Okuda staging of HCC. But as a parameter for radiation induced liver toxicity, it is yet to be validated (Table 4).

	0	1	2
Child Pugh stage	A	B	C
Tumor morphology	Unimodular & extension <50%	Multinodular & extension <50%	Massive or extension >50%
AFP	<400	>400
Portal vein thrombosis	−	+

Table 4.

CLIP scoring system.

4.3 ALBI score

ALBI score is a discriminatory method of assessing liver function in HCC with values of only albumin and bilirubin [11]. Validation of ALBI score as a tool in radiation toxicity assessment is undecided, but retrospective evidence indicates similar performance as with the CP score [12].

4.4 GRETCH score

The Groupe d’Etude et de Traitement du Carcinome Hépatocellulaire (GRETCH) score uses objective measures including bilirubin, alkaline phosphatase, AFP along with performance status and portal obstruction to predict survival outcomes. This prognostic system did not prove superior to other currently utilized scoring system and is not widely used world over [13] (Table 5).

Weight	0	1	2	3
Karnofsky index	>80			<80
Serum Bilirubin (umol/L)	<50			>50
Serum ALP	<2x ULN		>2x ULN
Serum AFP (ug/L)	<35		>35
Portal obstruction	−		+

Table 5.

GRETCH Scoring system.

4.5 AFP score

AFP is a well-established tumor marker for diagnosis of HCC that is detected in approximately 39–65% of HCC patients. AFP level normalization in a previously elevated patient within 3 months after SBRT is a prognostic surrogate for OS and PFS in patients with small HCC [14]. It is also useful in follow up of patients to detect early recurrence because the AFP level is related to the tumor activity. AFP stage for each prognostic group show clear survival differences (P < 0.0001), similar to the BCLC classification. However, survival differences among patient populations assigned to AFP stage B and C are not significant. In non-cirrhotic patients, the AFP staging system has a lower p-value than the BCLC classification.

4.6 PIVKA

Protein induced by vitamin K absence-II (PIVKA-II) is a potential screening marker for HCC and is an upcoming diagnostic tool that complements AFP [15]. Its role as a prognostic or predictive marker is yet to be determined.

Hepatocytes are involved in the synthesis of most coagulation factors, such as fibrinogen, prothrombin, factor V, VII, IX, X, XI, XII, as well as protein C, S, and antithrombin, whereas liver sinusoidal endothelial cells produce factor VIII and von Willebrand factor. Acute liver injury primarily decreases the vitamin K-dependent factors - prothrombin; factors VII, IX, and X.

5. Comparison between different evaluation systems

All these staging and scoring system have their own merits and demerits. Unfortunately, none of these scoring systems are validated in multiple prospective series. Hence, these systems are followed as per institutional preferences (Figure 2 and Table 6).

Figure 2.

Overlapping between different scoring systems.

	Okuda	CLIP	GRETCH
Child Pugh score		X
Ascitis	X
Albumin	X
Total bilirubin	X		X
Alkaline phosphatase			X
Alpha fetoprotein		X	X
Tumor size	X	X
Numbers of nodules		X
Portal vein thrombosis		X	X
Presence symptoms			X

Table 6.

Comparison between different scoring systems.

6. Factors responsible for RILD

6.1 Radiation dose and RILD

Liver is a fairly radio-sensitive organ. This has been evident from the pain control rates of 73–83% have been reported after RT for HCC [16, 17]. In the 1991 Emami report, the liver tolerance doses or TD 5/5 (dose expected to result in 5% complication rate in 5 years) were set as 50 Gy for one-third of the liver, 35 Gy for two-thirds of the liver, and 30 Gy for the whole liver [18]. Nevertheless, the primary liver tumors have not been irradiated with curative intent for a long period of time attributed to the conventional radiation portals practically including the whole organ.

With the advent of SBRT, very high doses can be delivered focally to the tumor, which are known to result in vascular injury and also an ablative effect on the tumor, in addition to the conventional DNA damage through dsDNA breaks. On the other hand, when these high doses of RT are being planned, one has to be extremely cautious regarding the precision and accuracy of the treatment. To account for inter and intra-fraction errors, various modalities like 4D CT, abdominal compression, voluntary breath hold, active breathing control and image-guidance during RT delivery can be used. The potential for tumoricidal doses to be delivered to focal HCC was first described by Dawson et al. at the University of Michigan by using an individualized dose allocation approach based on a normal tissue control probability (NTCP) calculation in 203 patients [19].

The Lyman–Kutcher–Burman (LKB)-NTCP model was used to assess dose-volume risk of RILD. The Lyman model assumes a sigmoid relationship between a dose of uniform radiation given to a volume of an organ and the chance of a complication occurring.

Various parameters have been looked into:

1. Effective volume (Veff): to allow volume-dose distribution comparisons between plans

2. TD50: tolerance dose associated with 50% chance of complication for uniform liver irradiation

3. m: steepness of dose response at TD50

4. n: defines the effect of the volume on a scale from zero to one [19].

Lausch et al. at the London Regional Cancer Program (LRCP), developed the logistic TCP model. They retrospectively reviewed 36 patients with HCC treated with median 4 Gy per fraction (range: 2–10 Gy) to a median cumulative dose of 52 Gy (range: 29–83 Gy) on a radiobiologically guided dose escalation protocol. The protocol called for prescribing the highest possible dose that met the constraint of keeping the estimated risk of RILD to <5%. They demonstrated that the D50 (dose that would result in a 50% LC) at 6 months was 53 Gy equivalent dose if given in 2 Gy fractions (EQD2). In contrast, the D50 for metastatic disease to the liver was 70 Gy EQD2 demonstrating that HCC is relatively radiosensitive compared to other tumor types, including colorectal carcinoma metastatic to the liver. The D90 was found to be 84 Gy EQD2 suggesting that increasing dose results in increased LC [20]. Jang et al. developed another logistic TCP model based on tumor size. They demonstrated that higher doses (cumulative and per fraction) are required to achieve the same TCP for larger lesions. For lesions <5 cm vs. lesions >5 cm, doses had to be escalated from 51 to 61 Gy in three fractions to achieve a 2-year LC of 90%. They have also reported that D50 was 62.9 Gy EQD2 (range: 58–69 Gy EQD2) [21].

Ohri et al. published another TCP model from data of 431 primary liver tumors and 290 liver metastases. The 1-, 2-, and 3-year actuarial local control rates after SBRT for primary liver tumors were 93%, 89%, and 86%, respectively. Lower 1- (90%), 2- (79%), and 3-year (76%) actuarial local control rates were observed for liver metastases (p = .011). Among patients treated with SBRT for primary liver tumors, there was no evidence that local control is influenced by BED within the range of schedules used. For liver metastases, on the other hand, outcomes were significantly better for lesions treated with BEDs exceeding 100 Gy₁₀ (3-year local control 93%) than for those treated with BEDs of ≤100 Gy₁₀ (3-year local control 65%, P < .001) [22].

In 2010, Quantitative Analysis of Normal Tissue Effects in the Clinic (QUANTEC) reported recommendations that mean normal liver dose should be <18 Gy for baseline CP-A patients and < 6 Gy for those with CP-B, for a 6-fraction SBRT regimen (Table 7) [3, 28].

Author	n	Selection	Dose	Radiation technique	Liver dose parameters	CP status (%)	Toxicity
Dawson et al. 2002 [19]	203	Unresectable intrahepatic cancer (HCC, cholangiocarcinoma, liver mets)	52.8 (range: 24–90)	3D-CRT	Median: 32.0 Gy (Range 14.9–44.0); LKB NTCP Median: 0.05 (Range 0.00–0.46)	—	RILD (n = 19, 9%) without RILD (n = 184, 91%)
Xi et al. 2013 [23]	41	HCC with marcovascular invasion	36 Gy (range, 30–48 Gy) in six fractions	SBRT using VMAT	Mean dose ≤18 Gy	Only CP class A	No Grade 4/5 toxicity
Andolino et al. 2011 [24]	60	HCC	CP A: 30–48/3 CP B: 24–48/5	SBRT	CP A: 1/3rd of the uninvolved liver was restricted to ≤10 Gy, and ≥ 500 cc of uninvolved liver received <7Gy. CP B: 1/3rd of the uninvolved liver was restricted to ≤18 Gy, and ≥ 500 cc of uninvolved liver received <12 Gy.	CP A: 36 CP B: 24	36.7%; 20% CP Progression 17 patients had grade 2 at baseline (out of 21 with grade ≥ 2 toxicity)
Tse et al. 2008 [25]	31	HCC and intrahepatic cholangiocarcinoma	36.0 Gy (24.0 to 54.0 Gy) in 6 fractions	SBRT	LKB – NTCP model	Only CP class A	Grade 3 liver enzymes were seen in five patients (12%).
Son et al. 2010 [26]	47	HCC	30–39 Gy (median: 36 Gy)	SBRT	V20 ≤ 50%; Total liver volume receiving <18 Gray (Gy) of radiation should be >800 cm^3	CP A: 89%; CP B: 8%; CP C: 3%	33% had ≥ Grade 2 hepatic toxicity; 11% had progression of CP class
Mizumoto et al. 2012 [27]	259	HCC	66 GyE in 10 fractions to 77.0 GyE in 35 fractions	Proton Beam Therapy	Mean dose and V0–30 were identified as significant factors; preferred V0: 30%	CP A: 198; CP B: 58; CP C: 3	Change in CP score ≥ 2 in 11% at 12 months and 22% at 24 months

Table 7.

Summarizing various trials involving liver RT.

The dose recommendations for SBRT as per QUANTEC [3] for 5% or less risk of RILD are:

Mean normal liver dose (liver minus gross tumor volume).

<13 Gy for primary liver cancer, in three fractions.

<18 Gy for primary liver cancer, in six fractions.

<15 Gy for liver metastases, in three fractions.

<20 Gy for liver metastases, in six fractions.

<6 Gy for primary liver cancer, Child-Pugh B, in 4–6 Gy per fraction (for classic or non-classic RILD).

Critical volume model-based ≥700 mL of normal liver receives ≤15 Gy in three to five fractions.

6.2 Liver volume and RILD

Liver being comprised of hepatic lobules as functional subunits is a parallel organ. As a result, the mean dose and a critical volume being spared of high dose is of significance rather than the Dmax. The University of Colorado phase 1 clinical trial of SBRT for liver metastases described the importance of the liver volume spared, that is, the ‘critical volume model,’ a concept akin to surgical sparing of the future liver remnant. They have estimated that a typical normal liver volume is approximately 2000 mL and specified that a minimum volume of 700 mL or 35% of normal liver should remain uninjured by SBRT i.e. at least 700 mL of normal liver (entire liver minus cumulative GTV) had to receive at total dose less than 15 Gy [29]. This critical volume concept has also been applied to patients with HCC. Dyk et al. retrospectively analyzed 46 patients, of which 91% are CP-A status, treated with liver SBRT for either metastatic or primary liver malignancies and found the liver volume at 25 Gy (V25) > 32% was associated with CP-class progression on Univariate analysis [30]. Son et al. retrospective review of 47 patients with HCC, of which 68% are CP-A status and showed the volume of normal liver receiving <18 Gy should be >800 cc to avoid CP class progression on Multivariate analysis [26]. Since all these studies constitute predominantly Child A patients, if these dosimetric parameters can be applied to Child B or C still uncertain. Indiana University group have further performed a phase II trial and reported their toxicity data in CP-A (n = 38) and -B (n = 21) patients [31]. For a treatment regimen of 48 Gy in 3 fractions, CP-A patients were required to either limit the dose to 33% of the uninvolved liver (D33%) < 10 Gy and/or maintain the liver volume receiving <7 Gy to <500 cc. For a more conservative treatment regimen of 40 Gy in 5 fractions, CP-B7 patients had to meet constraints of D33% < 18 Gy and/or > 500 cc receiving <12 Gy. Dosimetric correlates were identified for grade 3 to 4 hepatic enzyme toxicity observed in 10.5% and 38.8% of CP-A and CP-B patients, respectively. However, the lower limit of the normal liver volume seems to vary between different races and ethnicities. Because heights and body weights vary so is the body surface area and so is the normal liver volume. Hence an absolute normal liver volume or its percentage to be spared may not be the optimal parameter to evaluate the liver function required for patients. The concept of body surface area (BSA) and Basal Metabolic Index (BMI) guided estimation of optimal liver volume need to be spared during SBRT treatment may be the future of liver SBRT program.

6.3 Type of radiation and RILD

The conventional techniques like 2D and 3D CRT led to more of classic RILD owing to the wide radiation portals. With the technological advancements like IMRT, robotic SBRT with tumor tracking high accuracy in radiation treatment became possible and the necessity for additional ITV margin has been eliminated. Sharp dose gradient helps to deliver higher dose to the target and spare normal liver. With real time image guidance high precision therapy, PTV margin can be cut down. Thus, high doses can be focused to the tumor with minimal margin. Although the incidence of RILD decreased, this may led to higher probability of non-classic RILD.

6.4 Co-morbidities and RILD

Cirrhosis: Background liver Cirrhosis plays a major role in development of toxicity. Cirrhotic patients are more prone to develop non-classic RILD than normal patients. Also, evaluation of the radiation induced changes turn out to be a tedious process because the baseline liver function would also have been abnormal. Radiologic differentiation between radiation induced changes and disease progression is also challenging.

Infective etiology: Infective etiology as such is not directly related to radiation induced toxicity, but again the background inflammatory picture and liver functional status play a role in diagnosis of RILD.

Re-irradiation: McDuff et al. analyzed 49 patients who received re-irradiation to liver. Mean interval from initial RT to first re-treatment was 411 days (range 61–1668 days). Mean BED2 (α/β = 10) were 76.93 and 77.60 for initial treatment and re-treatment, respectively. Mean BED2 (α/β = 10) were 76.93 and 77.60 for initial treatment and re-treatment, respectively. Only 1 patient (2%) met criteria for “non-classic” RILD demonstrating significant metabolic derangements in the absence of progressive disease. Another 6 patients exhibited metabolic derangements in the presence of progressive intrahepatic disease burden [32]. There have been case reports of safe and effective delivery of radiation to liver multiple times [33, 34, 35, 36]. Appropriately selected patients under expert care can undergo re-irradiation in safety [37].

Nutritional status: Baseline nutritional status determines the general health of the patient and ability of the body to repair the radiation insult. Prior to the therapy, nutritional assessment thru hemoglobin, albumin, lactate dehydrogenase levels and the necessary corrections are recommended.

Disease stage: Disease status indirectly plays a role in development of toxicities. Larger the disease, larger will be the irradiated area and higher are the chances of RILD.

7. Re-irradiation in liver tumors

Re-radiation in liver tumors are not common in clinical practice. There are only few published literature in this aspect and no standard consensus regarding dosage schedule. In most of the subsites, such as in head & neck cancer or cervical cancer, in re-irradiation setting there is usually reduction of total dose (BED). Treatment volume is limited and fractionation schedule modified depending upon ‘time to re-treat’. Irradiated volume also important in selection of fractionation schedule. Usually, in head & neck cancer 7 year time is considered ‘safe’ to re-challenge with full dose of radiation therapy. In case of re-radiation before that period, there is a reduction of dose depending upon the ‘time to re-treat’. Usually 15% dose ‘decay’ considered in 1st year after radiation therapy and then every year 10% ‘decay’ in dose. As the time gap between primary radiation therapy and re-irradiation increase, safer to deliver higher (adequate) dose of radiation therapy to the target. In re-radiation of liver tumors this standard practice is not followed. In fact, in few studies there are better results (OS) in patients treated with higher dose in re-radiation setting. Child Pugh Score and ‘time to re-treat’ are considered significant prognostic factors. There is no compromise in irradiated volume as well. Tolerance of liver is low, but fortunately in re-radiation setting, liver tolerates radiation comparatively better than other subsites. High dose radiation therapy work like thrombo-embolism, embolizing blood supply to a portion of liver and stimulating proliferating of hepatocytes from adjacent normal liver. Proliferating hepatocytes causes hypertrophy of the liver portion which is naive to radiation therapy. This proliferating hepatocytes replace the post-CK necrotic liver. Hence, the ‘new’ regenerated portion of liver tolerate better than previously treated liver. Different cytokines liberated from the necrosed liver tissue may also stimulate hypertrophy of liver. After RT, there is fibrosis as well, and this fibrosis may lead to shrinkage of liver volume. Post-CK, there is 50% regression of the involved liver due to radiation injury, on the other hand there is 320% compensatory hypertrophy of the contralateral liver lobe [2]. This phenomenon negates the implications of firbosis, and hypertrophy has more predominant impact. Shrinkage of liver volume is expected to be more with higher integral dose of radiation therapy. In few studies, there is transient reduction of liver volume of about 20% at 3 months post-CK. However, at one year follow up there is only 10% shrinkage compared to pre-treatment volume. Even after repeat CK, liver volume is mostly maintained due to compensatory hypertrophy. Most severe complication after re-radiation is radiation induced liver disease (RILD). It is a syndrome of ascites, elevated transaminase level, and anicteric hepatomegaly. Usually occurs in a proportion of patient after receiving whole liver doses of >30–35 Gy. However, retrospective series of partial liver radiation have demonstrated that liver tolerance not only depends upon the total dose of radiation therapy, but also on pre-treatment Child-Pugh score, viral load and volume of tumor as well. Partial liver may be safely treated with radiation if adequate liver volume is preserved. In re-radiation, as the hypertrophied liver is mostly radiation naive, re-radiation is possible with adequate dose in small volume recurrences [33, 38].

8. Fiducial related toxicity

As stereotactic radiosurgery (SRS) applications moved to extra-cranial sites, the primary challenge was that SRS technologies were initially designed to deliver very precise treatments for non-moving targets. Therefore, methods to compensate for respiratory motion like fluoroscopy, surrogate markers [34] (spirometry, fiducials), 4D-CT and dynamic MRI were developed. Owing to the differential degree of movement of liver antero-posteriorly and cranio-caudally, and also between the lobes of liver, internal fiducial markers are ideal for tumor tracking. For fiducial tracking and CT slice thickness of 0.625 mm–1.25 mm, the system accuracy has been shown to be 0.7 +/− 0.3 mm. Per cutaneous fiducial insertion can be done ultrasonography guided or CT- guided under sterile conditions by interventional radiologist. Being an invasive procedure, complications like pain, bleeding, pneumothorax can be seen. Some of them might require chest tube placement, paracentesis, embolization. The technique of using “sterile blood patch” post fiducial insertion to prevent pneumothorax is in use. The main factor to prevent these remain the technical expertise. Apart from the acute complications, there can be migration of fiducials within the liver, rarely extra-hepatic sites also. Hence radiation planning and delivery is recommended to be scheduled after an interval of 48–72 hours post fiducial insertion [37, 39]. Park SH et al. retrospectively reviewed 101 patients with USG guided intrahepatic fiducial placement. There were no major complications, although 12 patients (12%) developed minor complications. Technical success was achieved in 291 (97%) fiducial placement. Of 101 patients, in 72/101 patients (71%) fiducials placement was ideal. Marsico M et al. (n = 15) assessed how different types of markers affects the tracking accuracy of Cyberknife. Ohta K et al. reported (n = 18) success rate of 100% (18/18) for fiducial placement in liver tumors. Only one patient (6%) had mild pneumothorax. There was no gross migration after placement. Choi J-H et al. (n = 32) evaluated the safety and technical feasibility of endoscopic ultrasonography (EUS)-guided fiducial placement. 23/32 patients (91%) had successful placement and only One patient (3%) developed mild pancreatitis which subsided with supportive care. Kim JH et al. (n = 77) evaluated the safety and technical success rate of an USG guided fiducial marker implantation. 21% had minor complications. Abdominal pain was the most common complication(14%). Fiducial migration occurred in 5 patients (6.5%). Dutta et al. analyzed 108 fiducials placed in 36 patients. Post-fiducial pain score 0–1 in 26 (72%) and score 3–4 was in 2 (6%). Five (14%) admitted in ‘day-care’ (2 mild pneumothorax, 3 pain). One patient (3%) admitted for hemothorax and died. Fiducial placement complications are usually rare, less than 3% patient need admission or have decompensation (change of Child Pugh Score > 2) [37, 39, 40].

9. Methods of prevention of RILD

The primary factor to prevent RILD is the better technique of radiation. SBRT with motion management techniques and real time tumor tracking is the best technique that can be used. Respecting the liver special constraints like mean liver dose and sparing a critical volume of liver from dose spill are the subsequent critical factors. Patient related factors like co-morbidities, nutritional status has to assessed prior to starting the treatment and the required dietary corrections have to be made. Feng et al. evaluated the role of amifostine as a radio protector in dose-escalated whole liver radiation therapy [41]. The study included 23 patients and a maximum dose of 40 Gy was used. This was compared with previously treated patients by logistical regression model. It was observed that the use of amifostine increased the liver tolerance by 3.3 +/−1.1 Gy. Selenium and Vitamin E are also shown to reduce the incidence of RILD in animal models by reducing liver lipid peroxidation and maintaining the endogenous liver antioxidant defense [34].

10. Management of RILD

No established therapies for classic RILD exist. There are no specific guidelines for the management of RILD. Suggestions for use of anticoagulants and steroids have been made, but it is primarily supportive care and diuretics are often used for the ascites. Although a few patients may recover, ample fraction will eventually die of liver failure. Thus proper patient selection to prevent RILD is crucial.

11. Conclusion

Radiation induced liver injury is potentially hazardous complication. There is no definitive treatment and a proportion of patient may land up in gross decompensation. Usually supportive care, diuretics, albumin supplement, vitamin K replacement may be useful. Better case selection will avert incidence of RILD. Precise imaging, contouring, planning and respecting normal tissue constraints are critical. Radiation delivery with motion management and image guidance will allow delivery of higher dose and spare normal liver and hence will improve response to treatment and reduce RILD.