Hepatocellular carcinoma (HCC) is the third most common cause of cancer death worldwide with increasing mortality rate in many countries [1–3]. HCC usually develops in patients with liver disease that compromises liver function [4]. Many HCC patients suffer from decreased liver function, and sometimes liver failure is the cause of mortality without cancer progression [5]. Hence, the prognosis of patients with HCC is complex by the interplay of tumor burden and liver function [6,7]. HCC management is complex, as one should carefully assess not only the risks and benefits of treatment on tumor, but also its effect on liver function. HCC is also notorious for its high recurrence rate even after curative treatment for early-stage tumor [8]. Hence, while managing patients with HCC, one should consider possibility of recurrence and available therapeutic option at the time of recurrence. Liver transplantation (LT) is a highly effective treatment option for HCC [9], and can dramatically change the prognosis of patients. However, donor availability differs by patient and regions, making things more complex. Recent advances in HCC treatment provided multiple potentially efficacious treatments, but, randomized controlled trials comparing between treatments are largely limited [7]. Several HCC guidelines have been published to facilitate the selection of optimal management [6,7,10], however, these guidelines have both similarities and dissimilarities due to the geographic differences, available resources, and lack of high level of evidences [11], and are not followed well in real-life clinical practice [12].

Multidisciplinary approach is a strategy that can help cope with escalating complexity in health care [13]. Multidisciplinary tumor board (MDT) approach are defined as an alliance of all medical and health care professionals related to a specific tumor disease whose approach to cancer care is guided by their willingness to agree on evidence-based clinical decisions and to co-ordinate the delivery of care at all stages of the process, encouraging patients in turn to take an active role in their care [14]. MDT management has been suggested as a tool to resolve a complex situation, where unpredictability, paradox and unknowable things remain [13,15]. HCC management is very complex, and hence, management via MDT may be a valuable option to provide optimal level of patients care [7,16]. However, MDT with participation by experts from multiple academic disciplines requires additional medical resources and time per patient, which leads to significant healthcare burden and high cost [17]. Several previous studies suggested that MDT management for HCC may improve survival [18–24], yet is limited by analyzing small number of patients [18–22], comparing patients who were treated in different time period [18,19,22], including those only received MDT management [20,21,23,24], or including those who received curative treatment only [20]. Recently, Serper et al., reported that MDT management was associated with reduced mortality, by analyzing 3,988 patients who received care through the Veterans Administration in involving 128 centers in U.S. [25]. However, survival advantage from MDT management might have been from receiving HCC care in high-volume center, where MDT management was available. Therefore, whether MDT management is necessary despite requiring more medical resources remains to be determined. In this study, we assessed whether management via MDT leads to improved survival in HCC patients.

The baseline characteristics of 6,619 HCC patients (median age: 57.0 years, men: 5,287 [79.9%], hepatitis B: 5,029 [76.0%]) are summarized according to the MDT management in Table 1. Patients who had MDT management were older, and included a higher number of men, had more preserved liver function, less advanced tumor stage, lower levels of AFP and PIVKA-II levels. Proportions of the patients who underwent resection, ablation or LT as a first-line treatment were similar for those who were managed via MDT than those who did not (49.2% vs.50.3%), but more patients received transarterial chemoembolization or other treatments (50.7% vs. 43.2%), and less patients received the best supportive care (0.1% vs. 6.4%) for those who were managed via MDT. There was no difference in terms of etiology and LT during follow-up.

When compared, there was no significance difference in the proportion of patients receiving BCLC guideline recommended treatment between MDT and non-MDT group (54.7% vs. 53.7%, p = 0.58). However, when stratified according to tumor stage, less patients received BCLC guideline recommended treatment in BCLC stage 0 (69.1% vs. 83.1%, p = 0.001), BCLC stage A (58.1% vs. 65.4%, p = 0.005), and BCLC stage C (5.3% vs. 11.8%, p = 0.035) in MDT group, and more patients received BCLC guideline recommended treatment in BCLC stage B (78.0% vs. 69.0%) in MDT group.

During the median follow-up of 3.5 years (range: 0.1–12.4 years), 3,266 patients (49.3%) died. The 1-, 3- and 5-year OS rates were 77.8%, 60.7% and 51.9%, respectively. The survival rate was significantly higher in patients who were managed via MDT compared with those who were not (71.2% vs. 49.4% at 5-year, P < 0.001, Fig 1A), with an adjusted HR of 0.47 (95% confidence interval [CI]; 0.41–0.53). MDT management was an independent factor for OS along with sex, etiology, ALBI grade, BCLC stage, AFP level, PIVKA-II level, and initial treatment modality (Table 2).

In this study, we observed better OS in HCC patients who had MDT management. Those received MDT management had different baseline characteristics than those who did not. Those received MDT management showed more preserved liver function, earlier tumor stage, lower levels of serum tumor markers, managed at more recent years, and had received more active treatment. However, MDT management was an independent factor associated with better OS in multivariable-adjusted model in the entire cohort. Also, MDT management was associated with better OS in the matched cohort, where the key factors related to patient outcome were exactly matched between patients who were and were not managed via MDT. The benefit of MDT management was observed across most pre-defined subgroups; especially, it was greater in those with poor liver function (ALBI grade 2 or ALBI grade 3), intermediate or advanced BCLC stage (BCLC stage B or C), or high AFP levels (≥200 ng/ml).

Our data is in line with previous studies which showed improved survival by MDT management [18–24]. The strength of this data is the relatively large sample size that enabled us to create exactly matched cohort. The better long-term outcome by MDT management in this study cannot be explained by differences in the baseline characteristics. It has been suggested that MDT approach is more beneficial when things are more complicated [13]. Interestingly, we observed that MDT management was more beneficial under more complex clinical situations (poorer liver function, intermediate or advanced tumor stage or higher AFP levels).

There are several plausible explanations underlying the improved survival associated with MDT management. It has been reported that MDT management can alters the direction of patient care by changes in imaging interpretation, revisions in diagnosis, and changes in treatment recommendations [23,27]. MDT management can also shorten the interval from diagnosis to treatment [22], and increase the odds of receiving treatment for HCC [18]. In a cohort study based on data from a MDT meeting dedicated to HCC, non-adherence to the MDT recommendation was one of the negative prognostic factors for OS [20]. MDT management has another advantage as patient is being managed by a group of specialists from multiple fields. HCC treatment selection is significantly influenced by provider subspeciality [28], that is, when assessed by only one specialist, optimal treatment selection for individual patient might be biased by provider’s experience and preferentiality [24]. A balanced treatment selection is ensured when specialists in each field openly discuss the risks and benefits of each treatment option. In this study, we could not present how many patients had change in imaging interpretation, revisions in diagnosis and changes in treatment recommendations, and how many patients had adherent to the MDT recommendation, as we have not recorded in this data prospectively. Notably, there was slight difference in treatment allocation between MDT and non-MDT group, careful interpretation is needed as this is an observational cohort study. Proportion receiving curative modalities (resection, ablation and LT) was similar between those in MDT and non-MDT group (49.2% vs. 50.3%). However, those receiving non-curative modalities (TACE and others) were much higher in MDT than non-MDT group (50.7% vs. 43.2%) in overall cohort. In matched cohort, less patients received curative modalities (49.4% vs. 56.4%) while more patients received non-curative modalities (50.6% vs. 39.8%) in MDT group. When further stratified according to BCLC stage, less proportion of patients in BCLC 0/A received BCLC-guideline recommended therapy (resection, ablation or LT) in MDT group, more proportion of patients in BCLC B received BCLC-guideline recommended therapy (TACE) in MDT group, and less proportion of patients in BCLC C received BCLC-guideline recommended therapy (sorafenib) in MDT group. However, in our practice, MDT usually discuss more complicated cases (e.g., small tumor located in segment 1, near major vessels or hilar area, which ablation or resection is not feasible or difficult), which explains why less proportion of patients received curative treatment in BCLC 0 or A in MDT group. Although MDT care may resulted in the revision in diagnosis, management or treatment which translated into better outcome, this study cannot answer whether observed better survival is from MDT care itself, or from selection of better patients with different characteristics in MDT group. Prospective evaluations are needed to see whether MDT care can really improve patient outcome, and to find out exact reasons behind the better outcome.

MDT management is not free of costs. It requires additional medical resources and costs. To date, MDT management is not reimbursed by National Health Insurance in Korea. MDT management has been practiced for free of costs in our institution till now. We had a limit for 15 cases per week for MDT discussion, which covers only about 10% of HCC patients. Expanding our MDT program was very challenging for multiple reasons including team capacities and cost issues. By simply gathering multiple academic disciplines together and allowing them to have a timely and open discussion, the outcome of HCC patients can be improved. In institutions where multiple academic disciplines are involved in HCC care, potential benefit of MDT care can be easily applied, as characteristics of MDT care do not require special equipment, rather, it requires only time and space for MDT member to have open discussion. A health care policy that supports MDT management is required to expand and sustain MDT management for HCC patients. Of note, we noticed that MDT management was especially helpful for those with poor liver function, intermediate or advanced tumor stage or high AFP levels. Given the restricted medical resources and cost, MDT management might be prioritized for a certain group of patients according to local regulations, team capacities and cost-benefit strategies.

This study has some other limitations. First, this is a retrospective cohort study. The decision to present a patient for discussion at MDT meeting was solely at the discretion of the provider in charge of the patient, and was not randomly selected patients. Hence, unidentified factors might be present that resulted in a better outcome for patients who received MDT care, including selection bias. Second, MDT management was not provided for every treatment among patients who have received multiple HCC treatment, but only a median of 1.4 times (min–max: 1–4) per each patient. HCC is notorious for its high recurrence rates. MDT management during all treatment courses of the patients might further benefit patients, which also requires prospective evaluation. Third, in subgroup analysis, those with ALBI grade 3 was small to have enough statistical power (n = 20). Fourth, the study was performed in South Korea, where hepatitis B is a major cause of HCC [29], and living-donor LT is a popular mode of LT [9]. Thus, generalizability of our findings to other regions with different etiologies and resources remains to be determined. Lastly, we observed better survival in MDT group for patients initially treated with transarterial chemoembolization, while this was not observed for patients initially treated with resection. Subsequent MDT management and treatment may explain this, as patients who are treated with transarterial chemoembolization usually experience higher recurrence than those treated with resection. However, exact reason for this is not clear.

In conclusion, management of HCC patients via MDT was associated with better survival. Especially, the benefit of MDT management was greater in those with poor liver function, intermediate or advanced tumor stage or higher AFP levels. MDT management might be prioritized for these groups of patients, if medical resources are of concern. MDT care has a potential to improve outcome of HCC patients where clinical situations are highly complex. Prospective evaluations are warranted and healthcare policy that can support MDT management for HCC patients are urgently required.