David S. Lu, Surachate Siripongsakun, Jeong Kyong Lee, and Phillip M. Cheng contributed to the study concept and design, the acquisition of data, the analysis and interpretation of data, the drafting of the manuscript, and a critical revision of the manuscript. Sindy H. Wei and Jong Seok Lee contributed to the study concept and design, the acquisition of data, and the analysis and interpretation of data. Saman Sabounchi contributed to the acquisition of data and the analysis and interpretation of data. Steven Raman contributed to the study concept and design, the acquisition of data, and a critical revision of the manuscript. Myron J. Tong and Ronald W. Busuttil contributed to the study concept and design, the analysis and interpretation of data, and a critical revision of the manuscript. James Sayre contributed to the analysis and interpretation of data and a critical revision of the manuscript.
Complete tumor encapsulation on magnetic resonance imaging: A potentially useful imaging biomarker for better survival in solitary large hepatocellular carcinoma
Article first published online: 11 MAR 2013
Copyright © 2013 American Association for the Study of Liver Diseases
Volume 19, Issue 3, pages 283–291, March 2013
How to Cite
Lu, D. S., Siripongsakun, S., Kyong Lee, J., Wei, S. H., Cheng, P. M., Sabounchi, S., Lee, J. S., Raman, S., Tong, M. J., Busuttil, R. W. and Sayre, J. (2013), Complete tumor encapsulation on magnetic resonance imaging: A potentially useful imaging biomarker for better survival in solitary large hepatocellular carcinoma. Liver Transpl, 19: 283–291. doi: 10.1002/lt.23597
- Issue published online: 11 MAR 2013
- Article first published online: 11 MAR 2013
- Accepted manuscript online: 24 DEC 2012 05:19AM EST
- Manuscript Accepted: 13 DEC 2012
- Manuscript Received: 30 AUG 2012
The aim of this study was to determine the prognostic value of complete tumor encapsulation as visualized on magnetic resonance imaging (MRI) in patients with a solitary large hepatocellular carcinoma (HCC) beyond the Milan criteria for liver transplantation (LT). Between December 2000 and March 2011, 57 patients who had a solitary HCC exceeding 5 cm in diameter at the time of initial MRI before any treatment were identified. MRI images of the patients were independently reviewed by 2 experienced readers for the presence of complete tumoral encapsulation. The medical records of the patients were reviewed for an outcome analysis. Thirty of the 57 patients had completely encapsulated HCC according to MRI. There was excellent interobserver agreement between the 2 readers for the assessment of complete encapsulation (κ=0.86). Overall survival was significantly longer for patients with completely encapsulated HCC versus patients with incompletely or nonencapsulated tumors (P<0.001), and this included a subanalysis of 33 patients who received locoregional treatment (LRT; P=0.04). The presence of complete encapsulation was a strong predictor for survival in these patients according to both univariate [hazard ratio (HR)=0.24, 95% confidence interval (CI)=0.12–0.52, P<0.001] and multivariate analyses (HR=0.25, 95% CI=0.07–0.85, P=0.03). The rates of down-staging (P<0.001) and eventual LT (P=0.02) after LRT were also significantly higher in the patients with completely encapsulated tumors. In conclusion, complete tumor encapsulation on MRI is a potentially useful predictor for favorable biology in patients with a solitary large HCC. This new imaging biomarker may have a role in treatment selection for patients whose tumors exceed the Milan criteria size limits. Liver Transpl 19:283–291, 2013. © 2013 AASLD.
Barcelona Clinic Liver Cancer
Half-Fourier acquisition single-shot turbo spin-echo
hepatocellular carcinoma with complete encapsulation
hepatocellular carcinoma without complete encapsulation
hepatitis C virus
magnetic resonance imaging
orthotopic liver transplantation
transcatheter arterial chemoembolization.
Hepatocellular carcinoma (HCC) remains one of the most common hepatic malignancies worldwide. Liver transplantation (LT) has become a curative treatment option for HCC, and the Milan criteria have been accepted as a standard for selecting HCC patients who can benefit from LT. However, it has been noted that patients with tumors beyond the Milan criteria size limits may also be favorable candidates for LT; therefore, the Milan criteria may be too strict because they rely solely on the size and number of tumors and do not consider any other phenotypic features of HCC. As a result, several institutions have expanded tumor size limits or have provided new criteria other than size for predicting survival outcomes after LT.[1-7]
The presence of a fibrous capsule on pathology is a characteristic finding of classic HCC and has been identified as a favorable prognostic factor after tumor resection.[8-10] In the radiological detection of fibrous capsules, magnetic resonance imaging (MRI) has been shown to be more sensitive than computed tomography (CT). However, no study has evaluated the encapsulation of HCC on MRI as a predictor of survival for HCC patients. Therefore, the purpose of this study was to determine whether a complete tumor capsule on MRI could serve as an imaging biomarker for better survival outcomes for patients with a solitary large HCC beyond the Milan criteria.
PATIENTS AND METHODS
The institutional review board approved this retrospective study, and the need for signed informed consent was waived. Imaging studies of all patients who underwent initial MRI for HCC before any treatment between December 2000 and March 2011 at our medical center were reviewed. Patients who had a solitary tumor measuring more than 5 cm in its longest diameter along with the pathological verification of HCC or typical radiological features (ie, hyperenhancement during the arterial phase and hypoenhancement during the portal venous or late phase) were included in the study. Sixty-five patients met these inclusion criteria. However, 8 patients were excluded because of the poor quality of their MRI scans. Therefore, 57 patients (47 males and 10 females) became the cohort of the study. The medical records of the patients were reviewed for demographics, laboratory values, clinical courses, treatment outcomes, and survival.
Because of the long time frame of the study (10 years), a wide variety of MRI systems and imaging protocols (including study protocols performed at outside institutions at the time of presentation) were used. Only studies of adequate quality meeting minimal standards were included in the study. These consisted of multiphasic, gadolinium-enhanced MRI on 1.5- and 3.0-T scanners with body phase-array coils. Sequences included axial T2-weighted multishot and/or single-shot techniques, T1-weighted in-phase and opposed-phase gradient echo, and pre- and post-gadolinium contrast-enhanced T1-weighted 2-dimensional or 3-dimensional gradient echo with or without fat suppression. The postcontrast sequences consisted of multiphasic acquisitions during the late arterial phase, portal venous phase, and at least 1 additional delayed venous phase with a section thickness of 6 to 8 mm for 2-dimensional acquisition on earlier generation scanners and a thickness of 2.5 to 5 mm for 3-dimensional acquisition on later generation scanners.
Two readers (D.S.L. and J.K.L. with 19 and 10 years of experience, respectively, in the interpretation of MRI scans of the liver) retrospectively reviewed all MRI scans with a picture archiving and communication system.
Assessment of Image Quality
Both readers evaluated the overall MRI quality to determine whether tumor borders were clearly depicted for the presence or absence of a capsule. The decision to exclude any compromised MRI scans was based on the consensus of both readers.
Measurement of Tumor Size
The longest diameter of the tumor was measured twice manually with a caliper and a picture archiving and communication system by the 2 readers in consensus. The average of the measurements was used for the analysis.
Evaluation of Complete Encapsulation
The 2 readers independently assessed the presence or absence of complete encapsulation of known HCCs on MRI. The presence of complete encapsulation was decided qualitatively on the basis of the following criteria after a review of all available magnetic resonance sequences: a clearly defined enhancing, high-signal-intensity rim on postcontrast portal or delayed-phase T1-weighted images and/or a low-signal-intensity rim on T1-weighted and T2-weighted images encompassing the entire tumor circumference. The assessment took into account partial volume effects at margins tangential to the section plane as well as overall confidence based on multiple planes (if available). When the tumor was nodular in contour, the nodular portion also had to be encapsulated completely and in continuity with the capsule of the main body of the tumor (Fig. 1C). The interobserver agreement for the assessment of complete encapsulation was evaluated. Then, to resolve discrepancies, the 2 readers re-evaluated MRI scans together to reach a consensus. Patients were classified into 2 study groups on the basis of consensus: hepatocellular carcinoma with complete encapsulation (HCC-C) and hepatocellular carcinoma without complete encapsulation (HCC-NC).
Evaluation of Vascular Invasion
Macroscopic vascular invasion on MRI was decided by the 2 readers in consensus and was considered present if there was evidence of direct tumor extension into adjacent portal or hepatic veins.
Tumor Progression or Treatment Response
All available imaging follow-up studies were reviewed for the degree of tumor progression and/or treatment response after locoregional treatment (LRT). Available CT, MRI, and explant pathology reports were retrospectively reviewed for the degree of treatment response according to the modified Response Evaluation Criteria in Solid Tumors system. All patients who received LRT were evaluated for the ability to be down-staged from beyond the Milan criteria to within the Milan criteria, regardless of the pretreatment tumor size. For down-staging to be considered successful, the tumor had to remain within the Milan criteria, as confirmed by CT or MRI, for at least 3 months after the treatment.
The baseline patient and tumor characteristics were compared for the HCC-C and HCC-NC groups with chi-square and Student t tests. The interobserver agreement for the presence or absence of complete encapsulation surrounding HCC was evaluated with the κ statistic. The level of agreement was interpreted as follows: κ=0.00–0.20, poor agreement; κ=0.21–0.40, fair agreement; κ=0.41–0.60, moderate agreement; κ=0.61–0.80, good agreement; and κ=0.81–1.00, very good agreement.
The overall survival and the subgroup survival of patients with supportive care only or LRT [ie, local ablation, transcatheter arterial chemoembolization (TACE), and hepatic resection] were evaluated with the log-rank test statistic from the Kaplan-Meier method and were compared for the HCC-C and HCC-NC groups. Survival was defined as the time measured from the initial diagnosis of HCC to death, with the data censored at the time of the last known record or at LT (if it occurred). The number of patients whose tumors were down-staged and the number who subsequently underwent LT were compared for the HCC-C and HCC-NC groups with the chi-square and Fisher's Exact test, respectively.
Univariate and multivariate analyses with a Cox proportional hazards model were used to determine the ability of each variable to predict survival. The multivariate analysis was performed with the forward stepwise selection technique to identify the independently predictive variables with a liberal P<0.1 level of significance as a retention criterion. Hazard ratios (HRs), corresponding 95% confidence intervals (CIs), and P values were reported. P<0.05 was considered to indicate statistical significance. Statistical analyses were performed with the IBM SPSS version 20 software package (IBM, Inc., New York, NY).
The baseline patient and tumor characteristics are summarized in Table 1. The mean age at the time of the initial diagnosis was 60.75 years (range=17–85 years). The median tumor size was 6.2 cm. Thirty of the 57 patients had complete encapsulation encompassing their HCCs according to MRI (Fig. 2), and 27 patients did not (Fig. 3). Child-Pugh class A, lower levels of alpha-fetoprotein (AFP), an absence of vascular invasion, and earlier Barcelona Clinic Liver Cancer (BCLC) staging classification were more common in patients with complete encapsulation versus patients without complete encapsulation, and this was statistically significant (P<0.001–0.01; Table 1).
|HCC-C Group (n=30)||HCC-NC Group (n=27)||Total (n=57)||P Value|
|Sex [n (%)]||0.61|
|Male||24 (80)||23 (85.2)||47 (82.5)|
|Female||6 (20)||4 (14.8)||10 (17.5)|
|Etiology [n (%)]||0.80|
|HBV||9 (30)||6 (22.2)||15 (26.3)|
|HCV||15 (50)||15 (55.6)||30 (52.6)|
|Other||6 (20)||6 (22.2)||12 (21.1)|
|Child-Pugh class [n (%)]||0.01|
|A||24 (80)||11 (40.7)||35 (61.4)|
|B||4 (13.3)||11 (40.7)||15 (26.3)|
|C||2 (6.7)||5 (18.5)||7 (12.3)|
|Tumor grade [n (%)]||0.68|
|I||7/24 (29.2)||1/9 (11.1)||8/33 (24.2)|
|II||11/24 (45.8)||5/9 (55.6)||16/33 (48.5)|
|III||6/24 (25)||3/9 (33.3)||9/33 (27.3)|
|Median tumor size (cm)†||5.8 (5.3–7.1)||7.6 (5.6–11.0)||6.2 (5.3–8.9)||0.04|
|Median AFP level (ng/mL)†||10.4 (3.9–51.7)||142 (7.9–4070.0)||28 (5.35–195.5)||0.005|
|Vascular invasion [n (%)]||<0.001|
|Yes||1 (3.3)||14 (51.9)||15 (26.3)|
|No||29 (96.7)||13 (48.1)||42 (73.7)|
|Metastasis [n (%)]||0.18|
|Yes||1 (3.3)||4 (14.8)||5 (8.8)|
|No||29 (96.7)||23 (85.2)||52 (91.2)|
|BCLC stage [n (%)]††||<0.001|
|B||28 (93.3)||12 (44.4)||40 (70.2)|
|C||0 (0)||10 (37.0)||10 (17.5)|
|D||2 (6.7)||5 (18.5)||7 (12.3)|
The distributions of HCC-C and HCC-NC patients who received supportive care only, LRT, and orthotopic liver transplantation (OLT) after LRT are shown in Table 2. The LRT group included 33 patients without evidence of local vascular invasion or distant extrahepatic metastases. Twenty patients received only TACE, 10 patients received both TACE and radiofrequency ablation, 1 patient received only surgical resection, and 2 patients received all 3 treatments. Since the Food and Drug Administration approval of sorafenib, 2 patients, both in the LRT group, received this systemic agent. One patient had HCC-NC treated by TACE and radiofrequency ablation and received sorafenib for only 2 months before death; the other patient had HCC-C treated by TACE, received sorafenib for 4 months, and was still alive at the time of the study analysis.
|HCC-C Group (n=30)||HCC-NC Group (n=27)||Total||P Value|
|Supportive treatment only (n)||6||18||24|
|Successful down-staging||21 (12)||2 (1)||23 (13)||<0.001b|
|Failed down-staging||3 (2)||7 (0)||10 (2)|
|OLT after LRT (n/N)||14/24||1/9||15/33||0.02c|
The comparative rates of transplantation for the HCC-C and HCC-NC patients were 14/24 and 1/9, respectively, and the difference was statistically significant (P=0.02; Table 2). Because the receipt of OLT may be influenced by factors other than tumor biology, the comparative tumor response rates (defined by success or failure in down-staging a tumor within the Milan criteria, ie, ≤5 cm) are also shown. Twenty-one of 24 HCC-C patients and 2 of 9 HCC-NC patients were successfully down-staged, and this also significantly favored the completely encapsulated tumors (P<0.001). Three of the down-staged patients eventually developed HCC recurrence, and other patients either were never evaluated for or had other morbidities that precluded OLT.
Kaplan-Meier estimates of overall survival and subgroup survival for patients with supportive care or LRT (censored at the time of transplantation) are shown in Fig. 4A-C. The HCC-C patients survived longer than the HCC-NC patients with a median overall survival of 26 months versus 6 months (P<0.001; Fig. 4A). In the supportive treatment group, there was a trend favoring the HCC-C patients (Fig. 4B), but in patients who received LRT, the difference significantly favored the HCC-C patients with a median survival of 50 months versus 10 months for the HCC-NC patients (P=0.04; Fig. 4C).
The single HCC-NC patient who was down-staged and then underwent LT died at 38 months from recurrent HCC. In contrast, the 14 HCC-C patients who underwent LT had an overall 5-year survival rate of 84% (Fig. 4D), with tumor recurrence occurring in only 1 patient (who died at 20 months). Detailed records of explant pathology were available for all except 1 patient, and these showed microvascular invasion in both patients who died with tumor recurrence. However, microvascular invasion was also present at explant in 3 of the other HCC-C patients, who were still alive 50, 81, and 99 months after transplantation.
On MRI, 29 of 30 patients with complete encapsulation showed an enhancing, high-signal-intensity rim surrounding their HCC on postcontrast T1-weighted images with or without a low-signal-intensity rim on precontrast T1-weighted and T2-weighted images. One patient showed a distinct low-signal-intensity rim on both precontrast T1-weighted and T2-weighted images but without a discrete enhancing rim on contrast-enhanced images. There was very good interobserver agreement between the 2 readers for the assessment of complete encapsulation with a κ value of 0.86 (95% CI=0.72–0.97). Macroscopic invasion of portal or hepatic veins was present in 15 patients (Table 1).
A univariate analysis of all variables revealed that the receipt of LRT (HR=0.13, 95% CI=0.06–0.29, P<0.001) and the presence of a complete capsule (HR=0.24, 95% CI=0.12–0.52, P<0.001) were significant predictors for survival. The significant predictors for mortality were Child-Pugh class B/C (HR=3.12, 95% CI=1.98–4.91, P<0.001), an increasing BCLC stage (HR=3.28, 95% CI=2.13–5.07, P<0.001), the presence of vascular invasion (HR=5.27, 95% CI=2.58–10.69, P<0.001), the presence of metastasis (HR=6.02, 95% CI=2.18–16.63, P=0.001), and the AFP level per log10 increase (HR=1.41, 95% CI=1.11–1.79, P=0.01; Table 3).
|HR||95% CI||P Value|
|Sex (male versus female)||1.404||0.54–3.64||0.49|
|Liver disease etiology (HCV versus non-HCV)||1.51||0.97–2.4||0.07|
|Child-Pugh class (B/C versus A)||3.12||1.98–4.91||<0.001|
|AFP (log10 ng/mL)ab||1.41||1.11–1.79||0.01|
|LRT versus supportive treatment||0.13||0.06–0.29||<0.001|
In a multivariate analysis, the presence of a capsule was an independent predictor of better survival with a 75% mortality risk reduction (HR=0.25, 95% CI=0.07–0.85, P=0.03). Increasing age (HR=1.06, P=0.01) and BCLC staging (HR=5.26, P=0.001) were also independent predictors of mortality (Table 4).
LT is an accepted treatment modality for patients with HCC and has excellent survival outcomes for select patients. The Milan criteria (a solitary HCC with a diameter≤5 cm or as many as 3 lesions with each diameter≤3 cm and no macroscopic vascular invasion or extrahepatic disease) are now widely used as the standard selection criteria for LT. Although the adoption of the Milan criteria is associated with excellent survival outcomes for HCC patients, there is a growing body of literature supporting the expansion or revision of the Milan criteria. On the basis of this observation, several institutions have recently proposed modified selection criteria for LT in patients with HCC; these include an expansion of the size or number limits for tumors in combination with an allowance for tumors to be down-staged into the Milan criteria by LRT.[2, 3, 5, 7, 13-17] Other centers have introduced new factors or biomarkers to help predict outcomes after LT, such as the serum AFP level, PIVKA-II (a protein induced by vitamin K absence or antagonist) level, and the tumor histological grade.[2, 4, 6]
Interestingly, among all the imaging features that are identified at the time of HCC diagnosis, other than obvious and macroscopically visible vascular invasion, only the size and number of tumors have been used for prognostication. Therefore, it would be valuable to have additional reliable imaging biomarkers for appropriate initial patient selection without additional tests or treatment challenges. A commonly recognized distinguishing feature among HCC nodules is the presence or absence of a capsule on imaging. We have long observed that some patients with very large tumors completely surrounded by well-defined capsules have had surprisingly long survival despite the lack of any treatment. On pathological studies, the presence of a fibrous capsule has long been identified as a favorable prognostic factor after tumor resection.[8-10] Such a barrier may serve to prevent or slow the invasion of HCC into the host liver or simply may be a marker of an effective host response to contain the tumor.[18, 19] To our knowledge, there has been no prior study addressing the predictive value of imaging-detected complete encapsulation of large HCCs with respect to survival. Our results confirm our hypothesis that the presence of complete encapsulation of a tumor on MRI can serve as an important imaging biomarker of favorable biology and may help in the selection of patients for therapy.
One of the goals for any patient with a liver-confined tumor beyond the Milan criteria is to ultimately undergo LT with or without down-staging therapy.[15-17, 20-22] In this study, complete tumor encapsulation at the time of diagnosis was predictive of a statistically significant higher rate of achieving both down-staging and LT in comparison with incompletely encapsulated HCCs. The concern that some of these patients may have undetected extrahepatic spread is allayed by the posttransplant outcomes. The 5-year survival for the 14 patients who had completely encapsulated solitary HCCs more than 5 cm in diameter was 84%, which was comparable to or higher than the average reported rate of posttransplant survival in all patients with HCC (most within the Milan criteria at the initial presentation).[23-26]
According to the univariate analysis, not surprisingly, well-known and long-established independent predictors for the survival of HCC patients, including Child-Pugh classification, AFP levels, macrovascular invasion, and extrahepatic metastasis, remained significantly associated with survival. Our multivariate analysis is also the first demonstration that the presence of a complete capsule on MRI is the single most powerful preoperative imaging-based predictor of survival for patients with a solitary HCC greater than 5 cm. When this factor is taken into account, it overwhelms other well-known tumor-based factors, including size and AFP. In fact, 2 of the 14 encapsulated tumors that were successfully down-staged (with subsequent transplantation and excellent posttransplant survival) exceeded 8 cm at presentation, which is the current upper limit for single tumors eligible for down-staging and a possible Model for End-Stage Liver Disease exception based on the modified University of California San Francisco down-staging criteria. The results of the current study, therefore, raise the possibility that the addition of this new imaging biomarker may improve the appropriate selection of potential LT candidates who have solitary tumors beyond the Milan criteria.
Interestingly, there has been considerable controversy in the literature regarding the role of LRT in HCC beyond the Milan criteria for down-staging to LT.[16, 17, 28, 29] Millonig et al., for example, showed no survival benefit of LRT in patients beyond the Milan criteria, whereas Yao and others[17, 29] showed excellent posttransplant survival rates. The findings of our study offer an explanation for this disparity: the different studies reported in the literature likely included tumors of heterogeneous biological behavior, one important indicator of which may be the presence or absence of complete encapsulation.
When a tumor appears encapsulated on MRI, there is some controversy about whether histologically there is a true fibrous capsule or a pseudocapsule consisting of compressed fibrovascular tissue and sinusoids. Regardless of the underlying histological nature, only the imaging identification of such a rim surrounding the tumor can potentially be a practical predictive tool. We also chose to classify the tumors only as completely encapsulated or incompletely encapsulated to enhance reproducibility. With this cutoff, the interobserver agreement was excellent in our study, and trained abdominal radiologists should be able to make the distinction between complete and incomplete encapsulation readily from MRI scans of sufficient quality.
In this study, we used only MRI for distinguishing between patients with and without complete tumor encapsulation. Such a capsule is also often readily discerned on contrast-enhanced CT. However, because the capsule is best demonstrated and often is seen only on delayed dynamic phase images and because not all CT scans are performed with adequately delayed phases (eg, single- or dual-phase CT), the overall CT detection of tumor capsules is less reliable than dynamic MRI detection. Also, capsule enhancement is more intense and is, therefore, more reliably visualized on gadolinium-enhanced MRI versus contrast-enhanced CT. Nevertheless, if complete tumor encapsulation is present on a good CT study, it should be expected to be similarly predictive of a favorable tumor biology.
There were several limitations to this study. First, it was restricted by its retrospective nature. Second, the overall number of patients was limited, although this sample currently represents, to the best of our knowledge, the largest cohort of solitary large HCCs with MRI and adequate follow-up and outcomes. Third, the population of patients was heterogeneous with respect to the type of treatment. To overcome this heterogeneity, we considered TACE, ablation, and hepatic resection to be in the same category of LRT, and the survival was censored at the time of LT to eliminate the overwhelming confounding effects of transplantation. Also, subgroup analyses were performed to assess separately the effects of LRT versus supportive care alone. Finally, we studied only solitary tumors larger than 5 cm. It remains to be determined whether this imaging biomarker has the same predictive value for tumors less or equal to 5 cm and whether it applies to multiple tumors.
In summary, the presence of complete tumor encapsulation on MRI is a potentially useful imaging biomarker associated with improved survival outcomes for patients with solitary large HCCs exceeding 5 cm in diameter whether they are untreated or treated by LRT and/or LT. Further studies should be conducted to validate these findings, and consideration should be given to incorporating this imaging biomarker into prognostication and treatment selection for patients with HCC, including future modifications of current LT selection criteria.