Abstract
Assessing the prognosis of patients with hepatocellular carcinoma (HCC) by the number and size of tumors is sometimes difficult. The main purpose of the study was to evaluate the prognostic value of total tumor volume (TTV), which combines the two factors, in patients with HCC who underwent liver resection. We retrospectively reviewed 521 HCC patients from January 2001 to December 2008 in our center. Patients were categorized using the tertiles of TTV. The prognostic value of TTV was assessed. With a median follow-up of 116 months, the 1-, 3-, and 5-year overall survival (OS) rates of the patients were 93.1 , 69.9, and 46.3 %, respectively. OS was significantly differed by TTV tertile groups, and higher TTV was associated with shorter OS (P < 0.001). Multivariate analysis revealed that TTV was an independent prognostic factor for OS. Larger TTV was significantly associated with higher alpha-fetoprotein level, presence of macrovascular invasion, multiple tumor lesions, larger tumor size, and advanced tumor stages (all P < 0.05). Within the first and second tertiles of TTV (TTV ≤ 73.5 cm3), no significant differences in OS were detected in patients within and beyond Milan criteria (P = 0.183). TTV-based Cancer of the Liver Italian Program (CLIP) score gained the lowest Akaike information criterion value, the highest χ 2 value of likelihood ratio test, and the highest C-index among the tested staging systems. Our results suggested that TTV is a good indicator of tumor burden in patients with HCC. Further studies are warranted to validate the prognostic value of TTV.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
Hepatocellular carcinoma (HCC) ranks as the fifth most common malignant cancer and the third most frequent cause of cancer leading death worldwide [1, 2]. Chiefly, due to the high prevalence of hepatitis virus infection and alcoholic liver disease, the disease burden of HCC is still quite tremendous [3, 4]. Along with the improvement in the population surveillance program and imaging modalities in the past decades, the proportion of patients at early or intermediate stages are expected to increase [5, 6].
Liver resection (LR) is considered as the mainstream treatment option for patients with well-preserved liver function, anatomic resectable lesions, and no evidence of distant metastasis [7–11]. Although liver transplantation removes the tumor lesions and the impaired liver simultaneously, its application is limited by the lack of liver grafts and by huge economic costs [12]. The prognosis of HCC patients largely depends on the interplay of the tumor burden and liver function reserve. Usually, the tumor burden is appraised by the number and size of the tumor nodules. The Milan criteria (solitary HCC diameter ≤ 5 cm or up to three nodules smaller than 3 cm), which was initially proposed by Mazzaferro et al. [13] in 1996, have been accepted as one of the standards selecting the candidates for liver transplantation. And, its prognostic efficacy was later validated in the realm of surgical resection [14]. However, a substantial proportion of HCC patients exceeding the Milan criteria have been found to survive longer than expected after curative liver resection [15]. It therefore seems reasonable to develop a tumor burden assessing parameter that gains optimal prognostic efficacy as well as further reduces unnecessary exclusions due to the strict application of the narrow selection criteria.
Recently, total tumor volume (TTV), which incorporates the number and size of the tumor nodules, has been proven to be a useful parameter to describe tumor burden in HCC patients awaiting liver transplantation [16–18]. However, scarce clinical evidences regarding the prognostic value of TTV in the LR-treated HCC patients have been reported. Lee et al. [19] compared the prognostic ability of Milan criteria and TTV in patients with TTV not more than 65.5 cm3. But, there have been no studies comprehensively evaluating the prognostic efficacy of TTV in HCC patients who underwent hepatectomy especially for patients with TTV > 65.5 cm3. We thus performed this retrospective analysis to evaluate the prognostic value of TTV in LR-treated patients.
Materials and methods
Patient selection
From January 2001 to December 2008, 539 patients with HCC as the only primary cancer who underwent curative liver resection (R0 resection) without previous history of treatments at the Department of Abdominal Surgical Oncology, Cancer Hospital, Chinese Academy of Medical Sciences, Beijing, People’s Republic of China, were identified. The diagnosis of HCC was determined by the evidence-based clinical guidelines and was confirmed by the pathological examination of the resected specimen. The negative resection margin was validated by the pathological examination. Of them, 18 patients were lost in the follow-up. Thus, finally, a total of 521 patients underwent R0 resection as the initial therapy for HCCs were taken into the study. Clinical data including patient demographics (e.g., age and sex), etiology of underlying liver disease, serum biochemistry, tumor number, tumor size, pathological reports, and liver function reserves were obtained from patients’ medical records. Information about tumor staging systems such as the seventh tumor-node-metastasis (TNM) staging of the American Joint Committee on Cancer (AJCC) [20], Cancer of the Liver Italian Program (CLIP) scoring system [21], Barcelona Clinic Liver Cancer (BCLC) staging system [22], and the Okuda staging system [23] for HCC were also collected. Patients’ informed consent was not required owing to the retrospective nature of the study.
Treatment
In our institution, the decision of curative intention LR was made by joint discussions. LR was usually considered in patients with (i) anatomically resectable tumors which were assessed by imaging examinations of the tumor size, tumor number, tumor location and vascular involvement; (ii) adequate liver function reserve (Child-Pugh grade A or B); and (iii) less than 25 % retention of indocyanine green 15 min after injection (ICGR-15).
Follow-up
Overall survival time of patients was calculated from the date of surgery to the date of last follow-up or death. The last date of follow-up was August 31, 2015. The follow-up was performed through face-to-face or telephone interview every 3 months in the first 2 years after operation and every 6 months thereafter or when tumor recurrence was highly suspected. At each follow-up visit, the information about physical examination, liver function tests, alpha-fetoprotein (AFP) level test, chest radiography, abdominal ultrasonography (US), contrast-enhanced computer tomography (CT) scans, and (or) liver magnetic resonance imaging (MRI) were obtained.
Definitions
The number and size of tumor nodules were measured by the CT scan in the preoperative evaluation. TTV was calculated as the sum of each tumor nodule volume; the volume of each tumor nodule was calculated as (4/3) × 3.14 × (maximum radius of the tumor nodule in cm)3 as previously described [16]. The patients were divided into three subgroups according to the tertiles of TTV (the first tertile <17.1 cm3, the second tertile 17.1–73.5 cm3, and the third tertile >73.5 cm3).
Statistical methods
Continuous variables were expressed as means with standard deviation (SD) or medians with range. Categorical variables were expressed as frequencies with percentages. For group comparisons, chi-squared test or Fisher’s exact test (categorical variables) and ANOVA test or Kruskal-Wallis test (continuous variables) were used to compare the differences between subgroups. Cumulative overall survival rates were appraised by the Kaplan-Meier method, and the differences were analyzed by the log rank tests. Significant factors identified in univariate analysis were subsequently enrolled in the multivariate Cox’s proportional hazard model. The likelihood ratio related to a Cox’s proportional hazard model was used to evaluate the homogeneity within categories of each system [24]. Discrimination for survival data was evaluated using the Harrell’s concordance index (C-index) [25]. The results of Cox’s regression were also presented using the Akaike information criterion (AIC) [26]. Cancer staging system with higher χ 2 value by the likelihood ratio test, higher C-index, and lower AIC value was considered to gain superior prognostic ability [26]. A two-tailed P < 0.05 was considered statistically significant. All statistical analyses were performed using SPSS 11.5 for Windows (SPSS Inc., Chicago, IL) and STATA version 10.0 (STATA Corp., College Station, TX, USA).
Results
The baseline characteristics of the patients
The baseline clinicopathological factors of the patients were summarized in Table 1. Among them, 453 (453/521, 86.9 %) were male patients and 68 (68/521, 13.1 %) were female patients. The mean age of the patients was 54.7 years old. Hepatitis B virus (HBV) was the dominant etiology of the underlying chronic liver disease in the study population (427/521, 81.9 %). Solitary tumor was detected in 442 (442/521, 84.8 %) patients. The maximum diameter of the tumor lesion was 4.7 ± 2.5 cm. The numbers of the patients at Child-Pugh class A and Child-Pugh class B were 509 (509/521, 97.7 %) and 12 (12/521, 2.3 %), respectively.
The distribution of the TTV
The TTV of the patients was 112.0 ± 222.4 cm3. The distribution of the patients with reference to the TTV, tumor number, and tumor size was shown in Fig. 1a, b. Adopting the cutoff values as 17.1 and 73.5 cm3, 174 (33.40 %) patients fell into the first tertile (<17.1 cm3), 172 (33.01 %) patients fell into the second tertile (17.1–73.5 cm3), and 175 (33.59 %) patients fell into the third tertile (>73.5 cm3). We compared the clinicopathological factors between the three groups. Patients with larger TTV tended to be featured with higher AFP level (P = 0.004; Table 2), presence of macrovascular invasion (P = 0.026; Table 2), multiple tumor nodules (P = 0.003; Table 2), larger tumor size (P < 0.001; Table 2), advanced TNM stage (P < 0.001; Table 2), advanced CLIP score (P < 0.001; Table 2), advanced BCLC stage (P < 0.001; Table 2), and advanced Okuda stage (P = 0.013; Table 2).
Prognostic factors for the patients’ survival outcome
With a median follow-up of 116 months (range, 80–175 months), the median overall survival (OS) of the patients was 58 months. The 1-, 3-, and 5-year overall survival rates were 93.1, 69.9, and 46.3 %, respectively. The median OS of patients in the first tertile was 69 months, which was significantly longer than that of patients in the second tertile as 60 months (P = 0.005; Table 3 and Fig. 2). And, the median OS of patients in the second tertile was also significantly longer than that of patients in the third tertile (median OS 60 vs. 40 months, P < 0.001; Table 3 and Fig. 2). In the univariate analysis, gender as male (P = 0.048; Table 3), presence of macrovascular invasion (P < 0.001; Table 3), albumin < 35 g/l (P < 0.001; Table 3), AFP ≥ 400 ng/ml (P = 0.018; Table 3), and larger TTV (P < 0.001; Table 3 and Fig. 2) were significantly associated with inferior survival outcome. These factors were taken into the subsequent Cox multivariate analysis. Compared with patients in the first tertile, the adjusted relative risk for mortality was 1.328 (95 % confidence interval (CI) 1.053–1.675, P = 0.016) for patients in the second tertile and 1.898 (95 % CI 1.502–2.397, P < 0.001) for patients in the third tertile. Besides, presence of macrovascular invasion (hazard ratio (HR) = 3.693, 95 % CI 2.140–6.374, P < 0.001; Table 3), albumin < 35 g/l (HR = 1.957, 95 % CI 1.364–2.809, P < 0.001; Table 3), and AFP ≥ 400 ng/ml (HR = 1.253, 95 % CI 1.016–1.546, P = 0.035; Table 3) were identified as the independent prognostic factors regarding the OS.
Comparison of the survival outcome within and beyond the Milan criteria in the first and second TTV tertile groups
Within the first and second tertiles of TTV (TTV ≤ 73.5 cm3), 312 patients were within Milan criteria and 34 patients were beyond Milan criteria. The median survival time of patients within and beyond the Milan criteria were 66 and 66 months, respectively. No significant difference in overall survival between the above two subgroups was detected (P = 0.183; Fig. 3).
Evaluation of TTV-based CLIP score
We further appraised the prognostic value of TTV by replacing the “tumor morphology” section of the original CLIP score with the TTV category. In the new system, TTV < 17.1 cm3 was assigned as score = 0, TTV within the range from 17.1 to 73.5 cm3 was assigned as score = 1, and TTV > 73.5 cm3 was assigned as score = 2 (Table 4). Statistically significant differences of OS were found in the pair-wise comparison across the TTV-CLIP scores. The median overall survival time across the TTV-CLIP scores 0–6 were 70.5, 65.5, 50, 41, 32, 20.5, and 15 months, respectively (Fig. 4a). We then compared the prognostic performance of TTV-CLIP score (Fig. 4a), CLIP score (Fig. 4b), TNM staging system (Fig. 4c), BCLC staging system (Fig. 4d), Okuda staging system (Fig. 4e), and Milan criteria (Fig. 4f). TTV-CLIP gained the lowest AIC (AIC = 4932.804), the highest likelihood ratio test (χ 2) as 34.99, and the highest C-index as 0.6126 among the seven prognostic scoring systems (Table 5).
Discussion
Few previous studies have extensively evaluated the prognostic value of TTV, which combined the information of tumor number as well as tumor size, in patients receiving liver resection. Our results suggested that patients with larger TTV gained a shorter overall survival than those with smaller TTV. Larger TTV was associated with elevated AFP level, multiple tumors, larger tumor size, presence of macroscopic vascular invasion, and advanced tumor stages. We further incorporated the TTV into the CLIP score and built the TTV-CLIP. The outperformance of TTV-CLIP score in predicting the clinical outcome further upheld the prognostic value of TTV in HCC patients.
Our results showed that TTV might provide a simplified way to describe the tumor burden and was also a potential parameter in the staging system for HCC. Compared with reporting the tumor burden by the tumor number and size, TTV has several advantages. TTV incorporates the number and size of tumor nodules into one continuous variable. Analyzing a single continuous index may be much simpler than analyzing the size and number of tumor nodule simultaneously. It was observed by someone that patients with multiple nodules but moderate tumor size may gain better clinical prognosis than those of patients with single large tumor [27]. It may be partially explained by the fact that the TTV of the former ones might be smaller than that of the latter ones. Regarding these, reporting the value of TTV instead of the size and number of tumor nodule may describe the tumor burden more accurately. We found that 34 patients in the first and second TTV tertiles were beyond the Milan criteria. The percentage of patients exceeding the Milan criteria but with TTV less than 73.5 cm3 was as high as 10.9 %, which was higher than the reported 6.5 % by Lee et al. [19] using the cutoff value as 65.5 cm3. The survival outcome of patients within Milan criteria was comparable with those of the patients beyond Milan criteria but with TTV ≤ 73.5 cm3. And, distinct difference in the OS was observed between patients with TTV > 73.5 cm3 and patients with TTV ≤ 73.5 cm3. The results suggested that the second tertile of TTV (73.5 cm3) in our study could be used as an expanded criteria selecting candidates receiving hepatectomy. Further studies are warranted to validate the prognostic efficacy of the criteria.
Patients with larger TTV predisposed to have AFP ≥ 400 ng/ml and macroscopic vascular invasion. Elevated AFP level and macroscopic vascular invasion were proved to be adverse prognostic indicators in our analysis, which were also consistent with previous studies [28, 29]. AFP level has long been regarded as being closely linked with the aggressive behavior of the tumor cells as well as the disease progression process [29]. And, several studies revealed that AFP concentrations often increase concomitantly with the augment of tumor burden [30]. Since TTV can be used to rating the tumor burden, it is quite possible that a larger TTV is expected to be more often associated with elevated AFP level. Besides, it is also assumed that larger tumors are often accompanied with the higher incidence of vascular invasion and presence of satellite nodules [27]. The close relationship between larger TTV and adverse clinicopathological features further defended the prognostic value of TTV.
CLIP score was initially derived from a retrospective cohort study of 435 patients and was then externally validated in a lot of studies [21, 31]. In our study, two of the components of CLIP score, including AFP level and presence of macroscopic vascular invasion, both obtained statistical significance in the multivariate analysis for OS. The original CLIP score has an inherent defect as the scale for “tumor morphology” is relatively subjective, without specific size criteria. Therefore, its objectivity and reliability in predicting outcomes may be somehow compromised [32]. We thus replaced the tumor morphology section with TTV in the modified CLIP score system. The TTV-based CLIP score gained a better prognostic ability than that of the original CLIP score, TNM staging system, Okuda staging system, BCLC staging system, TTV tertile category, and Milan criteria. Further external validation studies are warranted to establish the value of TTV-CLIP score.
This study has a few potential limitations. First of all, the retrospective nature of the study may be susceptible to the selection bias and recalling bias. Secondly, as HBV is the predominant etiology of HCC in China, the results of the study may have more implication for HBV-dominant areas. Our results should still be interpreted with caution in areas where other etiologies such as HCV, alcohol liver disease, and non-alcoholic fatty liver disease predominate. Thirdly, the calculation of TTV is based on the assumption that all tumor nodules are spherical. In fact, many tumor lesions were irregularly shaped. In order to enhance the applicability of our results in the surgeon’s preoperative evaluation, the value of radius was measured by the preoperative CT scan, which might be somehow different from the actual values. In addition, our study chiefly focused on patients who underwent liver resection. It should be reminded that before entering into the clinical practice, the prognostic values of TTV and TTV-CLIP scores need more external validations.
Conclusions
TTV is a good parameter measuring the tumor burden of HCC. Patients with TTV ≤ 73.5 cm3 may gain more survival benefits from liver resection. And, the TTV-CLIP score may provide a good prognostic performance for LR-treated HCC patients.
Abbreviations
- HCC:
-
Hepatocellular carcinoma
- LR:
-
Liver resection
- TTV:
-
Total tumor volume
- HBV:
-
Hepatitis B virus
- HCV:
-
Hepatitis C virus
- AFP:
-
Alpha-fetoprotein
- TNM:
-
Tumor-node-metastasis
- BCLC:
-
Barcelona Clinic Liver Cancer stage
- CLIP:
-
Cancer of the Liver Italian Program
- AIC:
-
Akaike information criterion
References
Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA: a Cancer J Clin. 2015;65(1):5–29.
Singal AG, El-Serag HB. Hepatocellular carcinoma from epidemiology to prevention: translating knowledge into practice. Clin Gastroenterol Hepatol. 2015;13(12):2140–51.
Forner A, Llovet JM, Bruix J. Hepatocellular carcinoma. Lancet. 2012;379(9822):1245–55.
de Martel C, Maucort-Boulch D, Plummer M, Franceschi S. World-wide relative contribution of hepatitis B and C viruses in hepatocellular carcinoma. Hepatology. 2015;62(4):1190–200.
Santi V, Buccione D, Di Micoli A, Fatti G, Frigerio M, Farinati F, et al. The changing scenario of hepatocellular carcinoma over the last two decades in Italy. J Hepatol. 2012;56(2):397–405.
Altekruse SF, McGlynn KA, Dickie LA, Kleiner DE. Hepatocellular carcinoma confirmation, treatment, and survival in surveillance, epidemiology, and end results registries, 1992–2008. Hepatology. 2012;55(2):476–82.
Bolondi L, Burroughs A, Dufour JF, Galle PR, Mazzaferro V, Piscaglia F, et al. Heterogeneity of patients with intermediate (BCLC B) hepatocellular carcinoma: proposal for a subclassification to facilitate treatment decisions. Semin Liver Dis. 2012;32(4):348–59.
Zhong JH, Ke Y, Gong WF, Xiang BD, Ma L, Ye XP, et al. Hepatic resection associated with good survival for selected patients with intermediate and advanced-stage hepatocellular carcinoma. Ann Surg. 2014;260(2):329–40.
Bruix J, Llovet JM. Prognostic prediction and treatment strategy in hepatocellular carcinoma. Hepatology. 2002;35(3):519–24.
Nathan H, Hyder O, Mayo SC, Hirose K, Wolfgang CL, Choti MA, et al. Surgical therapy for early hepatocellular carcinoma in the modern era: a 10-year SEER-Medicare analysis. Ann Surg. 2013;258(6):1022–7.
Fonseca AL, Cha CH. Hepatocellular carcinoma: a comprehensive overview of surgical therapy. J Surg Oncol. 2014;110(6):712–9.
Mazzaferro V, Lencioni R, Majno P. Early hepatocellular carcinoma on the procrustean bed of ablation, resection, and transplantation. Semin Liver Dis. 2014;34(4):415–26.
Mazzaferro V, Regalia E, Doci R, Andreola S, Pulvirenti A, Bozzetti F, et al. Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis. N Engl J Med. 1996;334(11):693–9.
Yamamoto J, Kosuge T, Saiura A, Sakamoto Y, Shimada K, Sano T, et al. Effectiveness of hepatic resection for early-stage hepatocellular carcinoma in cirrhotic patients: subgroup analysis according to Milan criteria. Jpn J Clin Oncol. 2007;37(4):287–95.
Ng KK, Vauthey JN, Pawlik TM, Lauwers GY, Regimbeau JM, Belghiti J, et al. Is hepatic resection for large or multinodular hepatocellular carcinoma justified? Results from a multi-institutional database. Ann Surg Oncol. 2005;12(5):364–73.
Toso C, Meeberg G, Hernandez-Alejandro R, Dufour JF, Marotta P, Majno P, et al. Total tumor volume and alpha-fetoprotein for selection of transplant candidates with hepatocellular carcinoma: a prospective validation. Hepatology. 2015;62(1):158–65.
Toso C, Trotter J, Wei A, Bigam DL, Shah S, Lancaster J, et al. Total tumor volume predicts risk of recurrence following liver transplantation in patients with hepatocellular carcinoma. Liver Transpl. 2008;14(8):1107–15.
Macaron C, Hanouneh IA, Lopez R, Aucejo F, Zein NN. Total tumor volume predicts recurrence of hepatocellular carcinoma after liver transplantation in patients beyond Milan or UCSF criteria. Transplant Proc. 2010;42(10):4585–92.
Lee YH, Hsia CY, Hsu CY, Huang YH, Lin HC, Huo TI. Total tumor volume is a better marker of tumor burden in hepatocellular carcinoma defined by the Milan criteria. World J Surg. 2013;37(6):1348–55.
Edge SB. American Joint Committee on Cancer. AJCC cancer staging manual. 7th ed. New York: Springer; 2010.
A new prognostic system for hepatocellular carcinoma: a retrospective study of 435 patients: the Cancer of the Liver Italian Program (CLIP) investigators. Hepatology. 1998;28(3):751–5.
Llovet JM, Bru C, Bruix J. Prognosis of hepatocellular carcinoma: the BCLC staging classification. Semin Liver Dis. 1999;19(3):329–38.
Okuda K, Ohtsuki T, Obata H, Tomimatsu M, Okazaki N, Hasegawa H, et al. Natural history of hepatocellular carcinoma and prognosis in relation to treatment. Study of 850 patients. Cancer. 1985;56(4):918–28.
Hosmer DW, Hosmer T, Le Cessie S, Lemeshow S. A comparison of goodness-of-fit tests for the logistic regression model. Stat Med. 1997;16(9):965–80.
Newson RB. Comparing the predictive powers of survival models using Harrell’s C or Somers’ D. Stata J. 2010;10(3):339–58.
Forster MR. Key concepts in model selection: performance and generalizability. J Math Psychol. 2000;44(1):205–31.
Jung YK, Jung CH, Seo YS, Kim JH, Kim TH, Yoo YJ et al. BCLC stage B is a better designation for single large hepatocellular carcinoma than BCLC stage A. Journal of gastroenterology and hepatology. 2015.
Nanashima A, Tobinaga S, Kunizaki M, Miuma S, Taura N, Takeshita H, et al. Strategy of treatment for hepatocellular carcinomas with vascular infiltration in patients undergoing hepatectomy. J Surg Oncol. 2010;101(7):557–63.
Bertino G, Neri S, Bruno CM, Ardiri AM, Calvagno GS, Malaguarnera M, et al. Diagnostic and prognostic value of alpha-fetoprotein, des-gamma-carboxy prothrombin and squamous cell carcinoma antigen immunoglobulin M complexes in hepatocellular carcinoma. Minerva Med. 2011;102(5):363–71.
Lee YH, Hsu CY, Huang YH, Su CW, Lin HC, Hsia CY, et al. Alpha-fetoprotein-to-total tumor volume ratio predicts post-operative tumor recurrence in hepatocellular carcinoma. J Gastrointest Surg. 2013;17(4):730–8.
Prospective validation of the CLIP score: a new prognostic system for patients with cirrhosis and hepatocellular carcinoma. The Cancer of the Liver Italian Program (CLIP) investigators. Hepatology. 2000;31(4):840–5.
Marrero JA, Kudo M, Bronowicki JP. The challenge of prognosis and staging for hepatocellular carcinoma. Oncologist. 2010;15 Suppl 4:23–33.
Acknowledgments
This study was funded by the National High-Tech R&D (863) Program of China (2015AA020408), the National Natural Science Foundation of China (81201967), the Beijing Natural Science Foundation (7144238), and Beijing Nova Program (No. 2009A69).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflicts of interest
None
Additional information
An erratum to this article is available at http://dx.doi.org/10.1007/s13277-016-4992-3.
Rights and permissions
About this article
Cite this article
Li, Mx., Zhao, H., Bi, Xy. et al. Total tumor volume predicts survival following liver resection in patients with hepatocellular carcinoma. Tumor Biol. 37, 9301–9310 (2016). https://doi.org/10.1007/s13277-016-4794-7
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13277-016-4794-7