Multifactorial Contributors to Diagnostic Delay in a High-Risk Patient with Hepatocellular Carcinoma: A Case Report with Clinical Lessons

Abstract

Background

Patients at high risk for primary liver cancer (PLC) are recommended to undergo PLC surveillance at least every six months. Alpha-fetoprotein (AFP) is a commonly used biomarker; however, approximately 15–30% of PLC cases remain AFP-negative throughout their course. Therefore, surveillance in high-risk populations should be based on liver imaging. Any newly detected or progressive lesion should be further characterized by contrast-enhanced computed tomography (CT), contrast-enhanced magnetic resonance imaging (MRI), or contrast-enhanced ultrasound, following guideline-based classification and follow-up recommendations.

Case Presentation

This case details a 45-year-old Chinese male with hepatitis B virus (HBV) infection, type 2 diabetes mellitus, and a family history of PLC. As a high-risk individual for PLC, he did not undergo regular PLC surveillance for three years. During opportunistic encounters for acute myocardial infarction and abdominal pain, the attending specialists failed to appreciate the progression of liver imaging findings. In addition, the COVID-19 pandemic may have further aggravated pre-existing difficulties with follow-up and access to care, contributing to a delayed diagnosis of PLC.

Conclusion

This case illustrates how a combination of physician-related, patient-related, and social-environmental factors contributed to surveillance failure and diagnostic delay in a high-risk patient with PLC. It reflects on why guideline-recommended interventions were not implemented at each healthcare encounter, how to correctly interpret AFP, and how to improve adherence to surveillance among high-risk populations for PLC. The insights from this case aim to inform systematic improvements in PLC monitoring for patients with chronic HBV infection.

Introduction

Primary liver cancer (PLC) is the third leading cause of cancer-related death worldwide.^1,2 In the Asia-Pacific region, chronic hepatitis B virus (HBV) infection accounts for 70–80% of PLC cases.³ Major international liver society guidelines consistently recommend PLC surveillance at least every six months for individuals at high risk of PLC.^2,4 Although AFP is a widely used, simple screening biomarker, its sensitivity is limited, with approximately 15–30% of PLC cases remaining alpha-fetoprotein (AFP) negative throughout the disease course.^5,6 This finding serves as a cautionary note to clinicians that AFP should not be used as a standalone rule-out test for PLC. Any newly detected or progressive lesion identified during surveillance should be further characterized using contrast-enhanced computed tomography (CT), contrast-enhanced magnetic resonance imaging (MRI), or contrast-enhanced ultrasound, with classification and follow-up interventions performed according to guideline recommendations.^2,7,8 This report presents a retrospective analysis of a case of diagnostic delay in an AFP-negative PLC, aiming to draw lessons from the experience.

Case Presentation

The patient was a 45-year-old Chinese male and self-employed businessman. In March 2011, following his father’s diagnosis of chronic hepatitis B and PLC, the patient was tested and found to have a hepatitis B virus deoxyribonucleic acid (HBV-DNA) level of 2.55×10⁶ IU/mL, leading to a diagnosis of “HBV carrier state”. He was initially treated with adefovir, which was later switched to entecavir for antiviral therapy. In August 2012, he was diagnosed with type 2 diabetes mellitus and initiated glucose-lowering treatment. Over the following years, he underwent regular follow-up for HBV-DNA, with results consistently <10 IU/mL, but his glycemic control remained poor. The clinical course from the first imaging detection of a liver lesion in December 2018 to the final diagnosis of PLC in January 2021 is described below.

December 2018

Prompted by his mother after his father’s death from PLC, the patient visited a general practice clinic. Laboratory investigations revealed the following: fasting blood glucose (FBG) 12.68 mmol/L, gamma-glutamyl transferase (GGT) 75 U/L, aspartate aminotransferase (AST) 18 U/L, alanine aminotransferase (ALT) 23 U/L, HBV-DNA ≤10 IU/mL, and AFP was normal (4.8 ng/mL) (Table 1). He underwent contrast-enhanced MRI for the first time, which revealed a non-rim arterial phase hyperenhancing nodule in segment VI of the liver, measuring approximately 10×12×10 mm, without portal venous phase washout or capsular enhancement. According to the liver imaging reporting and data system (LI-RADS) version (2018),⁹ the lesion was classified as LR-3 (intermediate probability of PLC) (Figure 1). And the general practitioner advised the patient to return to the hepatology clinic for follow-up imaging within 3–6 months.

Table 1.

Serial Measurements of AFP, HBV-DNA, Liver Enzymes, FBG, and Correlated Imaging Findings from December 2018 to January 2021

Parameter (Reference Range) Department		2018/12/18 General Practice	2019/02/14 Cardiology	2020/07/27 Endocrinology	2020/12/28 Emergency	2021/01/24 General Practice
AFP (≤7.00 ng/mL)		4.8	4.1	7.8	Missing	4.8
HBV-DNA (≤10 IU/mL)		≤10	≤10	≤10	Missing	≤10
AST (15–40 U/L)		18	320	28	Missing	51
ALT (9–50 U/L)		23	67	40	Missing	30
GGT (10–60 U/L)		75	136	223	Missing	412
FBG (3.6–6.2 mmol/L)		12.68	8.36	13.48	12.02*	10.51
Correlated imaging	Modality	Abdominal enhanced MRI	Missing	Abdominal enhanced CT	Emergency abdominal CT	Abdominal enhanced MRI
	Imaging finding (LI-RADS)	Segment VI (10×12×10 mm) LR-3	Missing	Segment VI (33×26×26 mm) LR-5	Segment VI (49×41 mm) PVTT	Right lobe (121×90×97 mm) Extensive PVTT

Note: *Random blood glucose.

Abbreviations: AFP, Alpha-fetoprotein; AST, Aspartate aminotransferase; ALT, Alanine aminotransferase; CT, Computed tomography; MRI, Magnetic resonance imaging; FBG, Fasting blood glucose; GGT, Gamma-glutamyl transferase; HBV-DNA, Hepatitis B virus deoxyribonucleic acid; LI-RADS, Liver imaging reporting and data system; PVTT, Portal vein tumor thromb.

Figure 1.

Contrast-enhanced MRI of the upper abdomen. An arterially enhancing nodule in segment VI of the liver (white arrowhead).

February 2019

The patient was admitted to the cardiology department due to acute myocardial infarction and underwent percutaneous coronary intervention with stent implantation. Routine laboratory tests showed FBG 8.36 mmol/L, GGT 136 U/L, ALT 67 U/L, and AST 320 U/L (attributed to myocardial infarction); AFP was normal (Table 1). The patient did not inform the cardiology team about the previously detected liver lesion, and the cardiologist did not review prior liver imaging findings. Repeat liver imaging had been recommended within 3 to 6 months after the December 2018 MRI and was not yet due when the acute cardiac event occurred approximately two months later; the planned surveillance process was therefore interrupted by the intervening acute cardiovascular event and subsequent clinical management. The patient was discharged after treatment for acute myocardial infarction only.

July 2020

During the domestic coronavirus disease 2019 (COVID-19) epidemic, the patient was admitted to the endocrinology department due to poor glycemic control (HbA1c 9.5%). Abnormal laboratory findings included FBG 13.48 mmol/L, and GGT 223 U/L; AFP was 7.8 ng/mL (Table 1). Contrast-enhanced CT revealed a poorly defined hypodense nodule in segment VI of the liver, which had enlarged to approximately 33×26×26 mm, with non-peripheral washout (ie, “fast washout”) in the portal venous phase and no definite capsular enhancement (Figure 2A and B). Compared with the baseline imaging in December 2018, the lesion had increased in maximum diameter from 12 mm to 33 mm (>50% increase) over an interval of approximately 19 months, with new nonperipheral washout. Because this interval exceeded the 6-month window for threshold growth in LI-RADS version 2018, the increase did not formally meet the threshold growth criterion; it is therefore described as marked longitudinal growth. Although the endocrinologist noted the abnormality, a hepatology consultation or multidisciplinary team discussion was not initiated. The medical record merely stated that the patient had been advised to follow up at a hepatology clinic. After discharge, the general practitioner failed to perform effective follow-up.

Figure 2.

Contrast-enhanced CT of the abdomen. The portal venous phase (A) and the delayed phase (B). A poorly defined hypodense nodule in segment VI of the liver, with non-peripheral washout in the portal venous phase and no definite capsular enhancement (white arrowhead).

December 2020

The patient presented to the emergency department with abdominal pain. However, the patient refused phlebotomy for laboratory testing; a random fingerstick capillary blood glucose level was 12.02 mmol/L (Table 1). Non-contrast abdominal CT showed a 49×41 mm hypodense mass in segment VI of the liver, with dilatation of the main portal vein and right branch accompanied by hypodense filling defects. These findings raised concern for portal vein thrombosis or possible tumor thrombus, although the nature of the thrombus could not be definitively characterized on a non-contrast examination. The emergency radiology report focused on the acute abdomen, noting only “impacted gallstone at the gallbladder neck”, without adequately emphasizing the liver mass or portal vein abnormality (Figure 3A and B). This imprecise interpretation may have contributed the emergency physician to focus on acute cholecystitis and impacted gallstone. The patient received anti-inflammatory and antispasmodic therapy, with symptom relief, and declined further inpatient evaluation.

Figure 3.

Abdominal CT of the patient. (A) A hypodense hepatic mass in segment VI of the liver (white arrowhead). (B) The main portal vein and its right branch were dilated and demonstrated hypodense filling defects (white arrowhead).

January 2021

The patient was admitted general practice department due to persistent abdominal pain, fatigue, anorexia, and weight loss. Abnormal laboratory findings included FBG 10.51 mmol/L, ALT 30 U/L, AST 51 U/L and GGT 412 U/L; AFP was normal (Table 1). Contrast-enhanced MRI revealed a large (121×90×97 mm) heterogenous lesion in the right lobe of the liver with arterial phase hyperenhancement, portal venous phase washout, restricted diffusion, and extensive portal vein tumor thrombus (Figure 4A and B). A final diagnosis of AFP-negative PLC, stage CNLC IIIa (China Liver Cancer Staging), was made. The patient had lost the opportunity for curative treatment and was offered palliative therapy with lenvatinib plus camrelizumab. Despite treatment, the patient ultimately died of PLC in May 2024, three years after diagnosis.

Figure 4.

Contrast-enhanced MRI of the upper abdomen demonstrated a large lesion in the right hepatic lobe. (A) The lesion showed arterial phase hyperenhancement (white arrowhead). (B) On the coronal plane portal venous phase, the lesion demonstrated washout and extensive portal vein tumor thrombus were identified (white arrowhead).

Discussion

This case analysis, driven by a chronological timeline, reveals multifaceted breakdowns in the PLC surveillance system, including insufficient implementation of PLC monitoring by general practitioners, failure of cross-specialty communication, impact of social-environmental factors, lack of standardized imaging reporting, poor patient adherence to follow-up, and misinterpretation of AFP at specific time points. The following dimensions are analyzed.

According to guidelines, an LR-3 lesion should be followed up with repeat or alternative imaging within 3–6 months.^9,10 However, after detecting the LR-3 nodule in December 2018, the general practitioner only advised the patient to attend follow-up without enrolling him into a chronic disease monitoring system. During three subsequent opportunistic clinical encounters (cardiology, endocrinology, and emergency department), each specialist focused exclusively on their own organ system and neglected the need for liver lesion follow-up. This exposes the fundamental flaws of cross-specialty information silos and an organ-centered rather than patient-centered care model, and underscores the failure to leverage the coordinating role of general practitioners in continuity of cancer surveillance for high-risk patients in primary care.^11,12 Therefor, Systematic failures in the surveillance system for high-risk PLC populations.

Contrast-enhanced CT performed in July 2020 demonstrated that the lesion had already acquired LR-5 imaging features (non-rim arterial phase hyperenhancement and non-peripheral washout with marked longitudinal growth from 12 mm to 33 mm over approximately 19 months). Although a lesion larger than 20 mm with non-rim arterial phase hyperenhancement and non-peripheral washout fulfills LI-RADS version 2018 criteria for LR-5 without the need for threshold growth, the radiologist did not apply the recommended LI-RADS categorization system at that time.⁹ For a lesion >2 cm in diameter in a high-risk individual, the presence of typical “fast-in and fast-out” features on at least one of four contrast-enhanced imaging modalities is sufficient for a clinical diagnosis of PLC without the need for liver biopsy.¹⁰ The failure of both radiologists and clinicians to recognize the diagnostic implication of an LR-5 lesion and to initiate a multidisciplinary team (MDT) discussion likely contributed substantially to the diagnostic delay in this patient.¹³

AFP exceeded the upper normal limit only once (7.8 ng/mL) among four measurements, and at the time of final PLC diagnosis, AFP was 4.8 ng/mL. The literature consistently shows that AFP-negative PLC accounts for 15–30% of cases.^5,6 A subsequent retrospective post-event discussion with the radiologist who interpreted the imaging, which provided retrospective information rather than contemporaneous documentation, suggested a tendency to exclude PLC on the basis of a normal AFP while overlooking the diagnostically more valuable typical imaging features of PLC. Retrospective accounts from post-event interviews indicated that clinicians repeatedly cited normal AFP values as a reason not to take MDT action and that they did not adequately explain or communicate with the patient; these accounts were not documented in contemporaneous medical records. Meanwhile, in a retrospective interview with the patient, the patient stated that he had always believed that PLC presented with elevated AFP because his father’s PLC had been AFP-positive, leading to the erroneous belief that a normal AFP rules out cancer—to the extent that he did not even undergo a liver ultrasound. Thus, AFP should only be considered an ancillary tool and must never be used as a basis to exclude PLC. This principle should be highlighted in every clinical decision-making encounter.

The patient declined regular follow-up. The emergency imaging examination performed in December 2020 for abdominal pain represented both an “accidental window” for early tumor detection and a “risk zone” for missed diagnosis or misdiagnosis. However, the patient refused the recommended further inpatient evaluation at that time. Review of the contemporaneous emergency radiology report, together with retrospective post-event interviews, suggested possible deficiencies in the emergency imaging assessment and reporting that may have impaired timely tumor detection, as well as a possible lack of effective physician–patient communication in the emergency setting. The contemporaneous medical records documented only the patient’s refusal of further inpatient evaluation; the underlying reasons were not recorded, nor were any attempts to explore the patient’s concerns, leverage family resources for risk communication, or establish a risk escalation or enhanced follow-up mechanism. The underlying reasons were therefore established retrospectively through post-event interviews with the patient and his family, which indicated that his reluctance was driven primarily by fear of a cancer diagnosis and subsequent procrastination. When a high-risk patient declines recommended follow-up, structured risk communication and shared decision-making are essential: clinicians should explore the patient’s concerns, correct misconceptions (including the belief that a normal AFP level excludes PLC), document the discussion, and agree on concrete options for continued surveillance and review.

During the COVID-19 pandemic, routine healthcare delivery was disrupted in many settings, with resources redirected toward epidemic control and reduced access to non-urgent outpatient care, and follow-up surveillance may therefore have been delayed or interrupted in some high-risk populations.^14,15 In this patient, however, the surveillance process had already been substantially interrupted before the major healthcare disruptions associated with the pandemic: after the acute myocardial infarction in February 2019, no liver imaging was performed for 16 months, although the patient had been informed that follow-up imaging was recommended within 3–6 months after the December 2018 MRI. He did not attend regular follow-up, received no systematic reminders, and consequently fell out of routine monitoring. COVID-19 may therefore have further aggravated the pre-existing difficulties with follow-up and access to care rather than being the primary cause of the diagnostic delay.

As a single case report, inherent bias exists, and generalizability of the conclusions may be limited. However, the data used are derived from original medical records, and the imaging features are directly verifiable. The problems exposed—fragmented specialty care, failure of standardized imaging reporting, and cognitive bias regarding AFP—have been reported in other healthcare settings, lending this case instructive value.

Conclusion

This case provides a multifactorial failure analysis based on a longitudinal timeline integrating laboratory data, imaging, and clinical decision points. We reflect on the guideline-recommended interventions that should have been taken but were not implemented at each medical treatment node. Learning from this case, implementing guideline recommendations, interpreting biomarkers dialectically, and establishing a multilayered safety net are essential to prevent similar tragic outcomes among high-risk populations for PLC.

Actionable Recommendations

1. For patients at high risk for PLC, a normal AFP level must never be used as a basis to exclude PLC.

2. An LR-3 lesion must trigger accelerated surveillance (imaging follow-up at 3–6 months) to preserve continuity of surveillance.

3. LR-4/LR-5 lesions should be promptly escalated through MDT discussion to ensure communication between specialties.

4. Standardized LI-RADS reporting should be widely adopted with each new study compared with previous imaging to support longitudinal review.

5. A traceable monitoring and tracking system for high-risk PLC patients should be established to prevent high-risk patients from being lost to follow-up.

6. When a patient declines recommended follow-up, structured risk communication and shared decision-making should be used, with the discussion and the agreed follow-up plan documented.

Funding Statement

This study was supported by Key discipline Foundation of Ningbo No.2 Hospital (No. 2023-Y10).

Abbreviations

AFP, Alpha-fetoprotein; ALT, Alanine aminotransferase; AST, Aspartate aminotransferase; CNLC, China Liver Cancer Staging; COVID-19, Coronavirus disease 2019; CT, Computed tomography; FBG, Fasting blood glucose; GGT, Gamma-glutamyl transferase; HBV, Hepatitis B virus; HBV-DNA, Hepatitis B virus deoxyribonucleic acid; LI-RADS, Liver imaging reporting and data system; MDT, Multidisciplinary team; MRI, Magnetic resonance imaging; PLC, Primary liver cancer.

Data Sharing Statement

The data presented in the study are included in the article, further inquiries can be directed to the corresponding author.

Ethics Approval

The study protocol was approved by the Medical Ethics Committee of Ningbo No.2 Hospital under the ethical code PJ-NBEY-KY-2025-235-01. Institutional approval was not required to publish the case details. This study was carried out according to the Declaration of Helsinki.

Consent for Publication

Written informed consent was obtained from the legally authorized representative(s)/next of kin of the deceased participant, consenting to the publication of any potentially identifiable images or data in this article. A copy of the written consent is available for review by the Editor-in-Chief of this journal.

Disclosure

The authors declare no potential competing interest.