The incidence of hepatocellular carcinoma (HCC) is increasing, both in Europe and worldwide. It is now the fifth most common cancer and represents a significant global public health burden. Unlike many other tumor types, diagnosis of HCC is frequently made without histological confirmation, based on non-invasive radiological criteria (NIRC) as recommended by both the European Association for the Study of the Liver (EASL) and the American Association for the Study of Liver Diseases (AASLD). These criteria can only be applied to cirrhotic patients and require a nodule of ≥1 cm showing the specific radiological hallmarks of HCC on CT or MRI, which differ from that of other liver tumors. The majority of studies investigating the diagnostic accuracy of these criteria are retrospective and focused on the detection of early stage disease. Despite this, they are frequently used for diagnosis in the advanced disease setting, and eligible patients often commence systemic therapy without a histological diagnosis.
In November 2017, NHS England mandated histological confirmation of HCC prior to commencing treatment with the tyrosine kinase inhibitor (TKI) sorafenib, which was at that time the only available first-line systemic therapy. This was based on the fact that histological confirmation of HCC was required for inclusion in the SHARP study, which lead to the European registration of sorafenib. This led to an overnight change in clinical practice across the UK and furthermore provided the opportunity to conduct a prospective study to investigate the diagnostic accuracy of NIRC in advanced disease and in a “real-life” setting. In this collaborative effort, we aimed to establish the safety and outcome of biopsy across the UK in patients with suspected advanced HCC.
Overall, 11 UK centres participated and data was collected for 418 patients identified at specialist multi-disciplinary team meetings and considered suitable for sorafenib therapy. In total, 361 patients met NHS England criteria for sorafenib therapy (performance status 0-2 and Child Pugh class A liver disease) and were therefore included in the primary analysis. Of these, only 164 (45%) patients fulfilled the NIRC meaning that even without the change in guidelines, 55% of patients in our audit would have required biopsy to confirm diagnosis according to international guidelines. We confirmed that biopsy is feasible in the advanced disease setting and took place in the majority of patients in whom there was no previous histology available (85%). Only a minority of patients (5.5%) could not proceed to biopsy due to either inaccessible disease or high risk factors such as coagulopathy.
The majority of biopsies confirmed an underlying diagnosis of HCC (93%), but a significant minority had alternative diagnoses including cholangiocarcinoma (CC) (4.3%), mixed HCC-CC (0.7%) and other malignancies such as breast and neuroendocrine tumors (2.3%). Considering those patients who met the NIRC, this translates to 65.4% sensitivity for NIRC to correctly identify HCC and an overall positive predictive value of 91.4%. This means that just under 1 in 10 patients with evidence of advanced radiological disease could be incorrectly classified as having HCC and receive inappropriate treatment for their underlying cancer. Furthermore, it means a similar proportion of patients could be recruited into HCC clinical trials without a true HCC diagnosis, thus influencing our therapies of the future.
In terms of the safety of biopsy in the advanced disease setting, there were only two reports of mild bleeding amongst the patients eligible for sorafenib and included in the primary analysis. Neither of these patients required any intervention or suffered any long-term sequelae. Amongst the entire cohort of 418 patients, there were a further two reports of bleeding. One of these was self-limiting but a second patient died from post-biopsy haemorrhage. Overall this equated to a 0.7% risk of mild bleeding amongst those patients who are eligible for systemic therapy with sorafenib, which is within the reported acceptable rates for liver biopsy.
This is the largest analysis of NIRC applied to advanced HCC to date, where histology has been used as a gold standard. One potential limitation of our work is that there was no central analysis of radiology and histology, meaning we are unable to differentiate between a true failure of NIRC to correctly diagnose HCC versus local misinterpretation of the imaging. We discussed this at the outset of this study, however, the purpose of this audit has always been to evaluate the efficacy of the NIRC outside the context of clinical trials and in the real-life clinical setting. By gathering local data across a number of UK centres we have calculated the diagnostic accuracy of NIRC in the hands of the specialists who will most frequently be using them.
Management of advanced HCC in the UK is mainly provided in tertiary centres and so is a relatively small community well known to eachother professionally. These professional networks were important in initiating this audit by allowing discussion of the main aims and intended methodologies from the outset. They also facilitated communication between sites with regards to the data collection and analysis and enabled timely feedback during the period of data collection. This audit has highlighted the importance of working collaboratively in order to access a large patient dataset representative of current UK practice. Successful cooperation across multiple sites has increased the impact of this audit which has implications for the management of advanced HCC patients.
An unintended consequence of the use of NIRC in HCC has been that there is limited tissue in advanced disease for basic science research and stratification of patients into clinical trials. Looking into the future, our recommendation for histological confirmation prior to commencing systemic therapy lends itself to the establishment of a biobank of tissue in HCC. This would be important in informing our understanding of cancer biology in advanced HCC and the development of new therapies.
Please sign in or register for FREE
If you are a registered user on Research Communities by Springer Nature, please sign in