KAHBPS 2015-04-25, Gyeongju, Korea Long-term outcome after resection of huge hepatocellular carcinoma ≥10 cm: Single-institution experience with 471 patients: Proposal of a prognostic prediction system Shin Hwang, Young-Joo Lee, Ki-Hun Kim, Chul-Soo Ahn, Deok-Bog Moon, Tae-Yong Ha, Gi-Won Song, Sung-Gyu Lee Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea
Huge HCC > 10 cm China > 20% of all HCC, Japan & Korea >10% of all HCC
Proportions of survival Overall and disease-free survival perioperative mortality: n=2 n=100 57% Proportions of survival Study period: 1997-2002 31% Overall Disease-free Postoperation months Lee, Hwang et al. World J Surg 2007
Background Zhang et al. World J Surg 2013 There is some noticeable survival difference according to tumor size, but it was not possible to identify any definite cutoff in tumor size. n=615 OS DFS OS DFS Cutoff in tumor size 2 cm 5 cm 3 cm 8 cm 4 cm 10 cm
Current issues If resectable, there is no limit in size for resection of HCC. Single HCC is regarded as T1 or T2 unless macroscopic vascular invasion is involved. The overall surgical outcome of huge HCCs is acceptable, but still worse than that of smaller HCCs.
Purpose To assess the long-term outcomes of resection for HCC ≥10 cm To establish a prognostic model for early recurrence, which will be useful for postoperative surveillance To validate the reliability of the prognostic model using an external cohort
Incidence of HCC > 10 cm 4148 HCC patients underwent resection during Jan. 2000 and Apr. 2012 - Exclusion of mixed pathology - 471 cases of HCC > 10 cm (11.4%) - Follow-up to Mar. 2014 (≥24 mos)
Flow of patients Surgical resection 471 patients Preoperative treatment (n=52) n=419 Surgical resection 1-month TACI (n=167) Perioperative mortality (n=8) Follow-up Tumor recurrence (n=346) No recurrence (n=117)
Preoperative treatment TACE RACE TACE + RTx/RFA RTx/RFA CTx n = 37 n = 8 n = 3 n = 4
Clinical features ------------------------------------------------------------------------------------------------------------- Age (yrs) 47.8±10.8 (range: 20 – 79) Sex (n) Male 375 Female 96 Primary liver disease (n) HBV 372 ALD 54 HCV 12 Others 33 Serum AFP (n=471) ≥100 ng/mL 304 (64.5%) Median 810 ng/mL (0.4-2700000) Serum PIVKA-II (n=231) ≥200 mAU/mL 304 (82.3%) Median 2400 mAU/mL (12-101073) ICG R15 (%) 13.5±8.2 Tumor diameter(median, cm) 13.0 (range, 10 – 26) Single tumor (n) 438 (93%) HBV, hepatitis B virus; HCV, hepatitis C virus; ALD, alcoholic liver disease; AFP, alpha-fetoprotein; PIVKA-II, Proteins Induced by Vitamin K Antagonist or Absence; ICG R15, indocyanine green retention test at 15 minutes.
Expression of AFP and PIVKA-II Linear scale Logarithmic scale PIVKA-II (mAU/mL) PIVKA-II (mAU/mL) AFP (ng/mL) AFP (ng/mL)
Ellipsoid tumor volume 10 cm-sized sphere = 500 mL 13 cm = 1000 mL 10 cm x 9 cm x 8 cm = 360 mL Mean diameter 13.6±3.1 cm 200 – 499 mL 500 – 999 mL ≥ 1000 mL n = 182 (38.6%) n = 175 (37.2%) n = 114 (24.2%)
Extents of resection -------------------------------------------------------------------------------------------------------------- Type of liver resection Systematic resection (n=429) (91.1%) Right hepatectomy ± S1 resection 251 Left hepatectomy ± S1 resection 69 Right anterior sectionectomy 23 Right posterior sectionectomy 31 Left lateral sectionectomy 14 Left medial sectionectomy 5 Central bisectionectomy 22 Right trisectionectomy 10 Left trisectionectomy 4 Non-systematic resection (n=42) (8.9%) Partial hepatectomy 42 Combined bile duct resection 8 (1.7%) Combined resection of adjacent organs (n) 28 (6.0%) Tumor thrombectomy Portal vein 47 (10.0%) Hepatic vein 4 (0.9%) Curative resection R0 resection 421 (89.4%) R1 resection 50 (10.6%) -------------------------------------------------------------------------------------------------------------
Extents of tumor ------------------------------------------------------------------------------------------------ Number of tumors (n) One 438 (93.0%) Two 19 (4.0%) ≥3 14 (3.0%) Simple nodular growth (n) 191 (40.6%) Microvascular invasion (n) 265 (56.3%) Macrovascular invasion (n) 49 (10.4%) Satellite nodules (n) 80 (17.0%) Bile duct invasion (n) 9 (1.9%) Capsule invasion (n) 86 (18.3%) Tumor necrosis (n) 380 (80.7%) Tumor differentiation (n) Most common Well: 168, Moderate: 194, Poor: 96 Worst Well: 88, Moderate: 261, Poor: 109 Regional lymph node metastasis (n) 12 (2.6%)
Long-term survival analysis Early recurrence prediction model External validation
Proportions of recurrence Cumulative HCC recurrence 76.0% 77.8% 72.5% Proportions of recurrence 62.2% n=471 Postoperation months
Proportions of survival Overall patient survival Perioperative mortality: n=8 (1.7%) 96.4% at 3 mos Proportions of survival 69.2% 46.9% 35.5% 18.8% Postoperation months
First recurrence timing and treatment -------------------------------------------------------------------First rec. site Treatment No. of patients ------------------------------------------------------------------- Liver (n=253) TACE 213 TACI 13 RFA 9 Resection 4 EBRT 1 PEIT 1 None 11 Lung (n=49) Chemotherapy 29 Resection 12 EBRT 3 None 5 Bone (n=7) EBRT 7 Adrenal (n=4) Adrenalectomy 3 EBRT 1 Thyroid (n=1) Thyroid lobectomy 1 Multiple recurrence (n=32) Specific treatment 22 None 10 -------------------------------------------------------------------TACE, transarterial chemoembolization; TACI, transarterial chemoinfusion; RFA, radiofrequency ablation; EBRT, external beam radiotherapy; PEIT, percutaneous ethanol injection therapy. Disease-free survival 4 mos 6 mos p=0.003 Surgical resection (n=20)
Overall survival respect to the first recurrence sites Disease-free survival Overall patient survival 20 mos 10 mos 21 mos p=0.009 p=0.003 + Post-recurrence survival 6 mos 12 mos 16 mos p=0.000
Univariate analyses for tumor recurrence and patient survival ---------------------------------------------------------------------------------------------------------------------------------------- Variables HCC recurrence Patient survival ------------------------------------- ------------------------------------- Median disease-free p-value Median overall p-value survival period (mos) survival period (mos) Serum AFP 0.041 0.000 < 100 ng/mL 12 53 ≥ 100 ng/mL 5 21 Serum PIVKA-II 0.542 0.443 < 200 mAU/mL 6 29 ≥ 200 mAU/mL 6 40 FDG-PET 0.000 0.000 Not hypermetabolic 36 > 5 years Hypermetabolic 5 25 Satellite nodule 0.000 0.001 Absent 8 37 Present 4 21 Microvascular invasion 0.000 0.000 Absent 16 58 Present 4 17 Macrovascular invasion 0.041 0.191 Absent 7 33 Present 4 20 Tumor volume 0.331 0.812 < 500 mL 8 27 50–999 mL 5 33 ≥ 1000 mL 7 29 Type of resection 0.079 0.014 R0 4 34 R1 7 11 --------------------------------------------------------------------------------------------------------------------------------------- AFP, alpha-fetoprotein; PIVKA-II, proteins induced by vitamin K antagonist or absence-II; FDG-PET, 2-18F-fluoro-2-deoxy-d-glucose positron emission tomography.
Multivariate analyses for tumor recurrence and patient survival ---------------------------------------------------------------------------------------------------------------- Variables HCC recurrence Patient survival ------------------------------------- -------------------------------------- Hazard 95% p-value Hazard 95% p-value ratio CI ratio CI Serum AFP ≥ 100 ng/mL 1.47 1.06–2.05 0.021 1.81 1.20–2.71 0.004 vs. < 100 ng/mL FDG-PET Hypermetabolic 1.79 1.02–3.15 0.033 2.18 1.01–4.77 0.041 vs. not hypermetabolic Satellite nodule Present 1.61 1.16–2.24 0.005 1.51 1.04–2.20 0.030 vs. absent Microvascular invasion Present 1.73 1.24–2.41 0.001 1.91 1.28–2.83 0.001 -------------------------------------------------------------------------------------------------------------- 95% CI, 95% confidence interval; AFP, alpha-fetoprotein; HCC, hepatocellular carcinoma; PIVKA-II, proteins induced by vitamin K antagonist or absence-II; FDG-PET, 2-18F-fluoro-2-deoxy-d-glucose positron emission tomography.
Long-term survival analysis Early recurrence prediction model External validation
Training cohort 4148 HCC patients underwent resection during Jan. 2000 and Apr. 2012 - 257 cases of HCC > 10 cm - Follow-up to Mar. 2014 (≥24 mos) - Meeting 4 factors: AFP, PET, pathology (satellite nodule, MVI)
Multiple regression for 3-month recurrence -------------------------------------------------------------------------------------------------------------------- Variables Beta SE of Beta B SE of B p-value AFP 0.149 0.062 0.153 0.064 0.017 (< vs. ≥ 100 ng/mL) FDG-PET 0.138 0.063 0.163 0.076 0.042 (Not hypermetabolic vs. hypermetabolic) Satellite nodule 0.151 0.061 0.173 0.069 0.014 (Absent vs. present) Microvascular invasion 0.148 0.064 0.148 0.064 0.021 Beta, standardized coefficient; B, unstandardized coefficient; SE, standard error; AFP, alpha-fetoprotein; FDG-PET, 2-18F-fluoro-2-deoxy-d-glucose positron emission tomography. Training cohort of 257 patients
4-risk factor prediction model - Training cohort of 257 patients - No. of risk factors : AFP, PET, satellite nodule, microvascular invasion Recurrence rate at 3 mos 63.2% 45.0% Percentage of recurrence 30.0% 12.1% 0% 0 1 2 3 4 No. of risk factors
Internal validation of 4-risk factor model for 12-month recurrence Tumor recurrence Patient survival 18.7% 30.3% 58.7% 79.0% 92.1% p=0.000 - Training cohort of 257 patients - No. of risk factors: AFP, PET, satellite nodule, microvascular invasion
Long-term survival analysis Early recurrence prediction model External validation
External validation cohort Study period: May 2012 – Feb. 2014 - 92 cases of HCC > 10 cm - Follow-up to Jan. 2015 (≥12 mos) - Patient profiles Age: 48.2 ± 11.3 (range, 23–73) Sex: male 71 (77.2%) HBV: 75 (81.5%) Single tumor: 84 (91.3%) Median tumor size: 12.8 cm R0 resection
External validation of 4-risk factor model for 12-month recurrence Tumor recurrence Patient survival 33.4% 31.6% 58.3% 78.3% 92.9% p=0.000 - Validation cohort of 92 patients - No. of risk factors: AFP, PET, satellite nodule, microvascular invasion The same discriminating power was obtained.
This study is under R2 minor revision in World Journal of Surgery. Conclusions Hepatic resection combined with active treatment of recurrence enhanced long-term survival in patients with HCC ≥10 cm. Our 4-factor risk prediction model appears to contribute to quantitative postoperative risk estimation for early HCC recurrence. Further validation is necessary with large patient populations in multicenter studies. This study is under R2 minor revision in World Journal of Surgery.