Influence of hepatocarcinoma surveillance programs on cirrhotic patient’s survival: a real life study.

Inclusion and exclusion criteria

The inclusion criteria was:

- All HCC diagnosed at our hospital, whatever the stage, between January 2012 and May 2018.

The exclusion criteria were:

- Space occupying hepatic lesions with a final diagnosis other than HCC (including other primary liver cancer, mixed lesions and non-HCC metastasic disease).

- HCC diagnosed outside the period under study.

Clinical and demographical characteristics

Different variables were retrospectively collected: epidemiological, demographic (age, sex), comorbidities (diabetes mellitus, arterial hypertension, HIV co-infection, other active neoplasms or in remission), aetiologies (HCV, HBV, MAFLD, alcohol, autoimmune, toxic, haemochromatosis , mixed or cryptogenic), liver function (Child-Pugh scale), laboratory variables (including alpha-fetoprotein at diagnosis) and clinical variables (ascites, encephalopathy, cirrhosis data and presence of portal hypertension). Regarding the tumour, the diagnostic method (radiological or histological), number and size of lesions were assessed. The condition of the patients was assessed using the ECOG-PS. Then, with the previous data, the BCLC stage was established. Finally, the treatment performed (liver resection, liver transplantation, radiofrequency, alcohol injection, chemoembolization, systemic therapies or best supportive care) and the response to it (remission, recurrence, survival and death), together with survival time, were reviewed.

Abusive alcohol consumption was defined as the intake of more than two standard drinks units in men and one in women. The presence of significant portal hypertension was defined by ultrasound data (portal diameter ≥ 14 mm, splenomegaly), laboratory tests (thrombocytopenia ruled out haematological causes), endoscopic data (presence of esophagogastric varices or portal hypertensive gastropathy) and haemodynamic data (hepatic venous pressure gradient ≥ 10 mmHg). Liver function was assessed using the Child-Pugh classification.

The patients in whom the diagnosis was considered to have been made during screening were those who were followed up in a specialized consultation with six-monthly ultrasound scans and for whom the tumour was first identified in one of these ultrasounds. AFP determination is not included as screening but is often monitored at the follow-up.

We define non-surveillance HCC diagnosis when HCC is detected because of hospitalization, when it is first detected by other radiological techniques or when patients don’t follow the screening properly (for example, one ultrasound per year).

Diagnosis, staging and response to treatment

The diagnosis of HCC was made following the clinical guidelines, where the lesions detected by ultrasound should present typical behavior for HCC on CT or dynamic MRI, that is, hyper-uptake in the arterial phase with lavage in the portal venous phase.8,9 In cases that did not meet these criteria or when the lesions were located on non-cirrhotic livers, a histological diagnosis was made by CT-guided biopsy.

Likewise, information regarding the tumour characteristics at diagnosis was collected according to the BCLC categories: single nodule < 2cm, single nodule ≥ 2 cm and ≤ 3 cm, single nodule > 3cm, up to 3 nodules < 3 cm; multinodular > 3 nodules or 2 or more nodules > 3 cm. Likewise, the presence or absence of extra-hepatic tumour involvement at the time of diagnosis was collected. With all this data, the patients were stratified according to the BCLC classification (0 or very early stage, A or early stage, B or intermediate stage, C or advanced stage, and D or terminal stage).

Regarding the treatment received, the decision was made with the support of a centre-based multidisciplinary Oncology Committee composed of hepatologists and gastroenterologists, surgeons in the hepatobiliary area, medical oncologists and radiotherapists, radiologists and interventional radiologists, specialists in nuclear medicine and pathologists. Patients with an indication for Liver Transplantation (LT) were referred to the reference transplant centre.

Parameters of response to treatment, complications, remission, recurrence, survival and death were collected; in addition to the time in months elapsed.

The response to treatment is assessed following mRECIST criteria:

-Complete response = disappearance of any intratumoural arterial enhancement in all typical intrahepatic target lesions and disappearance of all atypical intrahepatic target lesions and extrahepatic target lesions. Nodal lesions with short axis diameters regressed to -Partial response = at least a 30% decrease in the sum of diameters of the target lesions (including viable tumour diameters for typical intrahepatic target lesions and short axis diameters for nodal lesions), taking as reference the baseline sum of the longest diameters.

- Progressive disease = at least a 20% increase and an absolute increase of at least 5 mm in the sum of diameters of the target lesions (including viable tumour diameters for typical intrahepatic target lesions and short axis diameters for nodal lesions), taking as reference the nadir sum of diameters recorded since baseline.

- Stable disease = neither sufficient decrease to qualify for partial response nor sufficient increase to qualify for progressive disease.

- Not evaluable = at least 1 target lesion is not evaluable and the change in the sum of diameters of the measurable target lesions does not meet the criteria for progressive disease.

Statistical analysis

Statistical analysis was performed with the program SPSS version 23. Categorical variables were expressed as frequencies and percentages, while for continuous variables the mean and standard deviation or the median and interquartile range were used. For the comparison between epidemiological and clinical variables according to the screening group, the Chi-square or t-Student test was used, as appropriate.

The survival study was carried out using the Kaplan-Meier curves and the comparison between curves was assessed using the Mantel-Cox test (log-rank test).

Demographic and clinical characteristics at diagnosis (Table 1).

Of the 159 patients included in the study, 130 (81.8%) were male, with a mean age of 67.63 years, somewhat lower in patients diagnosed by the screening program (67.11 years vs 68.78 years), these differences being statistically significant. Regarding the most prevalent comorbidities at diagnosis, 50.3% had arterial hypertension, 34% had Type 2 diabetes mellitus, 3.8% had HIV co-infection, 8.8% had some active extra-hepatic neoplasia and 17.6% had suffered some other neoplasm that was in remission at the time of diagnosis, these data showing no significant differences.

Of the total sample, 147 patients (92.5%) had cirrhosis at diagnosis and 105 (66%) had evidence of portal hypertension. Cirrhotic patients were stratified according to the Child-Pugh classification: 87 patients (54.7%) were in stage A, 43 (27.7%) in stage B and 16 (10.1%) in stage C; 12 patients were not classified as they were not diagnosed with hepatic cirrhosis and 1 due to death before completing the study. When analysing the functional stage in the patients under the screening program, those included in this program were more frequently in early stages (stage A 69%, B 25% and C 2%) compared to those diagnosed outside the program (stage A 32 %, B 32% and C 24%), these differences being statistically significant (p <0.05).

The most frequent aetiology of liver disease was HCV infection (58 patients, 36.4%), followed by alcoholic liver disease (53 patients, 33.3%). However, in patients diagnosed outside the screening program, the most frequent aetiology was alcoholic cirrhosis (39%) and a relevant proportion of cases secondary to MAFLD (11.9%) was also found, resulting in statistically significant differences (p<0.05). The third most frequent aetiology was mixed (15 patients, 9.4%): mainly HCV and alcohol (12 patients), HBV and alcohol (one patient), and HBV and HCV co-infection together with alcohol (2 patients). MAFLD was observed in 11 patients (7%) and in 9 HBV infection as the sole aetiology. Other causes were autoimmune hepatitis, toxic hepatitis and haemochromatosis (one case for each, 0.6%). In 7 patients (4.6%), the aetiology of the liver disease could not be determined as the complete study was not concluded, while in another 3 it could not be established despite thorough diagnostic work-up, so they were considered cryptogenic (1.8%).

Diagnostic methods and characteristics of the tumour (Table 2).

The diagnosis was radiological in 126 individuals (79.2%) and histological in the rest. Of the total number of diagnosed patients, 100 (62.9%) originated from the screening program.

Regarding tumour size, 12.6% of patients had a single nodule < 2 cm at the time of diagnosis, 13.8% a single nodule ≥ 2 cm and < 3 cm and 38.4% a single nodule ≥ 3 cm. Five percent of patients had 3 nodules or less that were smaller than 3 cm, and 30.2% had multinodular involvement with > 3 nodules or ≥ 2 nodules ≥ 3 cm. Extra-hepatic involvement was detected in 18 patients (11.3%) and portal invasion in 25 (15.7%).

From the point of view of functional status, we found an ECOG 0 in 53.3%, ECOG 1 in 19.5%, ECOG 2 in 11.3%, ECOG 3 in 8.2% and an ECOG 4 in the remaining 7.5%.

According to the BCLC classification, a predominance of earlier stages was found in patients from the screening program (stage 0 9%, A 47%, B 17%, C 19% and D 3%), data very different from those presented by patients diagnosed outside that program (stage 0 1.7%, A 25.4%, B 10.2%, C 18.6% and D 44.1%), these differences being significant (p <0.05) (Figure 1).

Figura 1

BCLC staging distribution according to surveillance program

In 127 patients, the AFP level was determined at diagnosis, observing that 106 of these had values above the upper limit of normality (66.6%), with a median of 23.89 and an interquartile range of 273.54, without observing significant differences between patients undergoing screening or not (Table 2).

Therapy (Table 3)

Eleven patients were referred for liver transplantation (LT): one of them as first treatment, another after failed liver resection and the remaining 9 after previously receiving bridging therapy (4 chemoembolization (TACE), 4 radiofrequency and another combined TACE and RF therapy).

Hepatic resection (HR) was carried out in 21 patients, being the definitive treatment in 11 of them. Of the remaining 10, one was referred for transplantation, another required RF, 2 underwent TACE, 2 received systemic therapy, and the rest required a combination of various therapies.

RF was decided as initial treatment in 28 patients and TACE in 41 (in 9 of them as bridging therapy to transplantation). In 5 patients, the initial treatment was Percutaneous Ethanol Injection (PEI), of which 3 had to be completed with 3 sessions of TACE and in another with more sessions of PEI. During the course of the disease, 2 patients who started treatment with combined RF and/or TACE were subsequently treated with PEI.

Systemic treatment was indicated in 22 individuals (in 10 initially and in 12 at some point in the evolution). The systemic treatment used was Sorafenib in all cases. No other alternative therapies were used.

In 53 patients, symptomatic treatment was initially decided and in 11 more at some point in the evolution.

Overall, those patients who presented a BCLC stage that was favourable and initially eligible for therapies with better survival prospects (HT, RH, RF, and PEI) accounted for 40% of patients from screening versus 20% of those diagnosed outside of screening. In the same way, the patients who were indicated systemic or symptomatic treatment from the diagnosis were 7% and 16% respectively in the group from screening, 5% and 62.7% in patients not screened, these differences being statistically significant (p <0.05) (Table 3).

Regarding the complications recorded, one of the transplanted patients experienced acute rejection and a patient treated by liver resection developed post-surgical pneumoperitoneum that resulted in the death of the patient. As a consequence of TACE treatment, 3 patients developed post-chemoembolization syndrome, while one patient was identified for each of the following complications: fluid retention and ascites, hepatorenal syndrome, mild ischemic hepatitis, and pneumonia with associated pleural effusion. Of the patients treated with Sorafenib, 2 experienced diarrhoea, 1 toxicodermia, and 2 gastrointestinal intolerance.

Evolution and survival (Figure 2)

Figura 2

Survival analysis and sub-analysis. A) Overall survival analysis. B) Survival analysis depending on surveillance program follow-up. C) Survival analysis according to BCLC staging. D) Survival analysis in accordance with first line therapy received

The median follow-up obtained was 13 months (0-108), with loss of this in 4 cases. Of the 59 patients who received treatment with the best survival prospects, 29 (49%) achieved remission, 21 (36%) relapsed, and 9 (15%) had no response. Overall survival was found at 1, 2, 3 and 4 years of 60.3%, 47.3%, 37% and 27% respectively with a median survival of 22 months (95% CI 16.11-27.89) (Figure 2).

Evaluating the results according to the origin or not of the screening program, it was found that the patients coming from it had a median overall survival of 33 months (95% CI 20.25 - 45. 74) and a cumulative survival at 1, 2, 3 and 4 years of 72.3%, 55.5%, 44.2%, 34% versus an overall survival of 6 months (95% CI 1.97 - 10.02) and a cumulative survival of 1, 2, 3 and 4 of 36.7%, 30%, 22% and 0% in those diagnosed outside the program, this comparison being statistically significant (p <0.05)(Figure 2).

We performed a survival analysis based on tumour stage (BCLC classification at diagnosis). Survival was calculated by merging stage 0 and A for stability of the analysis; finding statistically significant differences when comparing these curves (Figure 2): a median of 50 months (95% CI 27.64-72.35) for early stage (0 and A), 20 months (95% CI 6.49 - 33.55) for stage B , 10 months (95% CI 4.91 - 15.08) for stage C and 2 months (95% CI 0.44 - 3.55) for stage D.

Survival analysis was performed based on demographic and tumour characteristics at diagnosis, some of these, in turn, forming part of the BCLC staging system. Regarding the characteristics of the tumour lesion, a median survival of 43 months (95% CI 20.41 - 65.58) was found for single nodules ≤ 3 cm, 24 months (95% CI 15.59 - 32.40) for single nodules > 3 cm, 22 months (95% CI 0 - 58.61) for lesions with less than 3 nodules and these < 3 cm and 9 months (4.98 - 13.01) for lesions with more than 3 nodules or these > 3 cm; the difference between these survival curves being statistically significant (p log-rank <0.05).

The survival related to the functional stage (Child-Pugh classification) was similar at diagnosis, with the median for stage A being 35 months (95% CI 21.43 - 48.56), stage B 10 months (95% CI 0 - 20.74) and 2 months stage C (0 - 6.75). Survival in patients with portal hypertension was significantly lower (median 16 months, 95% CI 7.55 - 24. 45) versus those who did not present it (median 46 months, 95% CI 31.60-60.39). Likewise, patients with portal tumour invasion at diagnosis showed a lower survival than the rest (median 3 months, 95% CI 0.72 - 5.27 vs 26 months, 95% CI 17.06 - 34.93). The comparison of the survival curves according to the ECOG classification also showed statistically significant differences with a median of 50 months (95% CI 27.41 - 72.59) for stage 0, 24 for stage 1 (95% CI 18.85 - 29.15), 6 for 2 (95% CI 1.81 - 10.18), 5 for 3 (95% CI 0 - 10.36) and 0 for 4.

Finally, the median survival rates obtained in our series, according to the initial treatment, were the following (Figure 2): 58 months (95% CI 44.43 - 72.71) in transplant patients (with bridge therapy or not), 45 months (95% CI 36.53 - 55. 08) for HR, 43 months (95% CI 29.2 - 56.79) in those treated with RF, 33 months (95% CI 16. 02-49.97) with PEI, 21 months (95% CI 9.84 - 32.15) for TACE, 12.5 months (95% CI 3.96 - 21.03) for systemic therapy and 5 months (95% CI 2.37 - 7.62) in the best supportive care treatment group ; the difference between the curves being statistically significant (p log-rank <0.05).

No statistically significant differences were found in survival based on the other epidemiological, clinical and diagnosis-related characteristics.