Exploring the MRI and Clinical Features of P53-Mutated Hepatocellular Carcinoma

Introduction

More than 90% of primary liver cancers are hepatocellular carcinoma (HCC) and are predisposed to occur in patients with underlying chronic liver disease.¹ HCC has a high mortality rate, and the survival rate of patients with advanced HCC is very low.² HCC has a poor prognosis mainly because it is prone to vascular infiltration, metastasis and recurrence.³ It is important for patients to be able to make preoperative predictions about the prognosis of hepatocellular carcinoma and to choose appropriate treatment options. Therefore, additional data about the pathogenesis of HCC are needed to investigate new and more effective treatments to improve the survival of HCC patients.

P53 is an important tumor suppressor gene involved in the development of hepatocellular carcinoma, and it is closely related to cell cycle arrest, cellular regulation and cellular senescence.⁴ While the wild-type P53 gene is a tumor suppressor gene, mutant P53 has oncogenic activity, this protein not only binds to wild-type P53 to form a complex and inhibit P53 gene function but also promotes abnormal cell proliferation and participates in tumor formation.⁵ The most commonly mutated genes in hepatocellular carcinoma are TERT (promoter), TP53, CTNNB1, AXIN1, ARID1A, and ARID2, whereas the chance of mutation occurring in other genes is less than 10%.⁶ Enrichment of TP53 inactivating mutations promotes the onset of HCC proliferation, including in clinically aggressive tumors with poor differentiation and frequent vascular invasion. The mutation rate of the P53 gene in HCC patients is 31.5%, and the expression up-regulation is 35.0%. Moreover, P53 expression can lead to a poor prognosis in HCC patients.⁷

Small HCCs with a diameter of ≤2 cm have low invasiveness, but their incidence is increasing annually, and the optimal treatment options are uncertain. Small-diameter (less than 2 cm) HCCs have high resection rates and low recurrence rates.⁸ Liver transplantation is the best therapeutic option for patients with unresectable HCC. The Milan criterion for liver transplantation is a single nodule ≤5 cm in length.⁹ Mazzaferro et al¹⁰ reported that among patients who had undergone liver transplantation for HCC, the 4-year recurrence-free survival rate was 83%, and the ability of patients to choose the right treatment helped improve their survival. Liver transplantation also has some limitations in that it mainly selects HCC that meets the Milan criteria, and overall survival remains low for out-of-criteria HCC and patients with high-risk HCC features (with microvascular invasion and poorly differentiated tumors).¹¹

P53-mutated HCC is a major subtype of HCC, and the ability to identify P53-mutated HCC in a noninvasive manner can help patients develop a personalized treatment plan. Several studies have shown that P53-mutated HCC is associated with dilated vasculature in the arterial phase and a decreased relative enhancement ratio in the hepatobiliary phase of gadoxetic acid–enhanced MRI;¹² however, imaging manifestations in the ancillary features of HCC (mosaic architecture, nodule-in-nodule architecture, etc.) have not been investigated, and their diameters have not been further categorized. Therefore, we investigated the imaging features of HCCs expressing P53 mutations on contrast-enhanced MR and explored whether there were differences in the imaging features of P53-mutated HCCs with different diameters.

Materials and Methods

Patient Selection

The study was approved by the institutional review board committees of Zhongshan Hospital, Fudan University (approval number: B2021–113R), and patient informed consent was waived. Our study was retrospective and therefore consent was waived. Our study covers patient data confidentiality and compliance with the Declaration of Helsinki. Between January 2020 and December 2020, a total of 910 consecutive patients were confirmed to have HCC by postoperative pathology with P53 immunohistochemistry and no extrahepatic metastases on preoperative examination. The inclusion criteria were (a) primary liver lesion without any previous treatment; (b) MRI examination within 30 days before the operation, for which the MRI scan was clear and suitable for diagnosis; (c) a single mass; and (d) a maximum lesion diameter between 10 mm and 100 mm. Finally, 566 patients were excluded for the following reasons: lack of clinical data (n = 15); history of previous treatment, including surgery and transarterial chemoembolization (TACE) therapy (n = 242); lack of MRI data available 1 month before the procedure or poor MRI images, which made it difficult to make a diagnosis (n = 32); two or more HCC lesions in the same liver (n = 178); and lesions with a maximum diameter less than 10 mm or more than 100 mm (n = 99), because the imaging signs of larger tumors are complex in presentation and affect the judgement of the results. Finally, 344 patients with HCC were enrolled in this study (Figure 1).

Figure 1 Patient selection process and exclusion criteria. Abbreviations: HCC, hepatocellular carcinoma; MRI, magnetic resonance imaging; LTD, the largest tumor diameter; P53(+), P53-mutated; P53(-), non-P53-mutated.

Histologic Diagnosis and Immunohistochemistry

Pathologic analyses, including immunohistochemistry, revealed the following features: tumor differentiation according to the Edmondson–Steiner grade and the presence of microvascular or macrovascular invasion. Immunohistochemistry for P53 was performed for all 344 patients with HCC, and then the patients were classified into P53-mutated HCC and non-P53-mutated HCC groups. P53 expression in tumor tissue was scored according to the percentage of cells exhibiting positive nuclear staining; P53 expression was defined as positive when 10% or more of the cells exhibited nuclear staining.¹³

MRI Technique

All patients were examined Gd-DTPA enhanced liver MRI. Patients were scanned using the following nine MRI scanners: a Magnetom Aera 1.5T imager (Siemens Healthcare, Erlangen, Germany) for 55 patients; a 3.0T UIHMR770 scanner (United Imaging Healthcare, Shanghai, China) for 102 patients; a Magnetom Avanto 1.5T imager (Siemens Healthcare, Erlangen, Germany) for 44 patients; a 1.5T UIHMR560 scanner (United Imaging Healthcare, Shanghai, China) for 76 patients; and a Magnetom Verio 3.0T MRI System (Siemens Healthcare, Erlangen, Germany) for 14 patients; a Prisma 3.0T MRI System (Siemens Healthcare, Erlangen, Germany) was used for 26 patients, and Ingenia CX3.0 MRI System (Philips Medical Systems, Best, the Netherlands) for 27 patients. Gd-DTPA was intravenously administered at a rate of 2 mL/s for a total dose of 0.1 mmol/kg. The routine MR imaging protocol included transverse T2-weighted breath-hold fat-suppressed fast spin‒echo sequence, T1-weighted breath-hold in-phase and opposed-phase gradient echo sequence, and free-breath diffusion-weighted imaging (DWI) with a transverse single-shot spin‒echo planar sequence (b value, 0, 50, and 500 s/mm²). The contrast-enhanced MR imaging protocol included pre- and postcontrast three-dimensional T1-weighted imaging in the arterial phase (20–30 s), portal venous phase (70–90 s) and delayed phase (160–180 s).

Imaging Analyses

Images were analysed by two radiologists (** and ** with 9 and 16 years of abdominal imaging analysis experience, respectively) retrospectively. If there was an inconsistency between the two radiologists, a consensus was reached through discussion. The following findings were assessed Gd-DTPA enhanced MRI (Figures 2–4): (a) enhancement patterns (hypoenhanced-isoenhanced, nonrim arterial phase hyperenhancement and rim enhancement); (b) washout patterns (no washout, nonperipheral washout, peripheral washout); (c) dilated vasculature, defined as thick branching vessels of similar contrast to arteries with a maximum diameter >2 mm; (d) delayed central enhancement, defined as progressive enhancement in the central region of the lesion in the delayed phase; (e) enhancing capsule, defined as an enhancing rim of tissue that appears during the portal venous and delayed phases; (f) corona enhancement, defined as periobservation enhancement in the late arterial or early portal venous phase; (g) mosaic architecture, defined as randomly distributed internal nodules or compartments on T2WI images; and (h) nodule-in-nodule architecture, defined as a mass with a smaller internal nodule within a larger nodule accompanied by marked enhancement in the arterial phase with washout in the portal phase. (i) A tumor in the vein was defined as an occluded vein adjacent to the intrahepatic mass. In addition, intratumoral hemorrhage, fat deposition, restriction diffusion status (yes or not), and hepatic capsule retraction were also assessed.

Figure 2 Images in a 40-year-old man with P53-mutated HCC; the serum level of AFP was 4.2 ng/mL. (A) T1-weighted image shows a 27mm hypointense mass. (B) T2-weighted image shows a hyperintense mass. (C) Axial arterial phase image shows a rim enhancement mass. (D) Delayed phase image shows delayed enhancement in the center of the mass.

Figure 3 Images in a 38-year-old man with P53-mutated HCC; the serum level of AFP was 3.6 ng/mL. (A) T1-weighted image shows a 17mm hypointense mass. (B) T2-weighted image shows a hyperintense mass. (C) Axial arterial phase image shows a hypervascular mass. (D) Delayed phase image shows a mass with enhancing capsule.

Figure 4 Images in a 50-year-old man with P53-mutated HCC; the serum level of AFP was 11.4 ng/mL. (A) T1-weighted image shows a 42mm hypointense mass. (B) T2-weighted image shows a hyperintense mass. (C) Axial arterial phase image shows a mass with nodule-in-nodule architecture and irregular peritumoral enhancement. (D) Delayed phase image shows nodule-in-nodule architecture washout and peritumoral slight hypointensity.

Clinical Data

Clinical information such as age, sex, background liver characteristics, liver functional parameters, and serum tumor marker levels of AFP, CA19–9, and CEA was collected.

Statistical Analysis

All statistical analyses were performed using SPSS 26.0. Inter-reader agreement was expressed by Cohen’s kappa coefficient. A kappa statistic of 0.81–1.0 was considered very good, 0.61–0.80 was considered good agreement, 0.41–0.60 was considered moderate agreement, 0.21–0.40 was considered fair agreement, and 0–0.20 was considered poor agreement. Continuous variables were compared using Student’s t test, and categorical variables were compared among different groups using the χ2 test or Fisher’s exact test. Parameters were analysed using univariate and multivariate logistic regression to determine whether they were independent risk factors for P53-mutated HCC. Univariate analyses were performed first, and only those parameters found to be statistically significant (p < 0.05) were used in stepwise multivariate logistic regression.

Results

Patient Clinical

In our study, a total of 344 patients were included, 196 cases (57.0%) were P53-mutated HCC and 148 cases (43.0%) were non-P53-mutated HCC, of which 54 cases had HCC with a maximum diameter of 10 mm to 20 mm, 33 cases (61.1%) were P53 mutated and 21 cases (38.9%) were non-mutated; 189 cases had HCC with a maximum diameter of 20 mm to 50 mm, 98 cases (51.8%) were P53 mutated and 91 cases (48.1%) were non-mutated; and 101 cases had HCC with a maximum diameter of 50 mm to 100 mm, 65 cases (64.4%) were P53 mutated and 36 cases (35.6%) were non-mutated.

There were significant differences in the largest tumor diameter, AFP level, Edmondson–Steiner grade and microvascular invasion between P53-mutated HCC patients and non-P53-mutated HCC patients (p = 0.018, p = 0.012, p < 0.001 and p = 0.034, respectively; Table 1). In patients with HCC with a diameter greater than or equal to 10mm and less than or equal to 20 mm, there were no significant differences in age, sex, diameter, etiology, liver function parameters, AFP, CA 19–9, CEA levels, Edmondson–Steiner grade, or microvascular invasion (p > 0.05; Table 2). In patients with HCC with a diameter greater than 20 mm and less than or equal to 50 mm, there were significant differences in the Edmondson–Steiner grade between P53-mutated HCC and non-P53-mutated HCC (p = 0.023; Table 3). Among patients with HCC with a diameter greater than 50 mm and less than or equal to 100 mm, P53-mutated HCC had greater maximum diameters and serum AFP levels than non-P53-mutated HCC (p = 0.011 and p = 0.008, respectively; Table 4).

Table 1 Clinical Characteristics HCC According to P53 (10mm ≤ LTD ≤ 100mm)

Table 2 Clinical Characteristics HCC According to P53 (10mm ≤LTD ≤ 20mm)

Table 3 Clinical Characteristics HCC According to P53 (20mm ˂ LTD ≤ 50mm)

Table 4 Clinical Characteristics HCC According to P53 (50mm ˂ LTD ≤ 100mm)

Imaging Features

The interobserver agreement for MRI features was good, with a kappa coefficient ranging from 0.704 to 0.958 (Table 5).

Table 5 Interobserver Agreement of MRI Imaging Features

There were significant differences in hepatic capsule retraction and corona enhancement between P53-mutated HCC patients and non-P53-mutated HCC patients (p = 0.047 and p = 0.021, respectively; Table 6). Among the groups with a diameter greater than or equal to 10mm and less than or equal to 20 mm, the P53-mutated group was more likely to have an enhancing capsule than the non-mutated group (p = 0.045), and hemorrhage, fat, rim enhancement, mosaic architecture, nodule-in-nodule architecture, and LI-M were more likely to occur in the P53-mutated group, however, there was no statistically significant difference in the p value (p ˃ 0.05; Table 7). In the 20 mm to 50 mm diameter group, the P53-mutated group had a higher incidence of corona enhancement and nodule-in-nodule architecture (p = 0.023 and p = 0.029; Table 8). In the groups with a diameter greater than 50 mm and less than or equal to 100 mm, hepatic capsule retraction, hemorrhage, fat, rim enhancement, dilated vasculature, delayed central enhancement, enhancing capsule, corona enhancement, and LI-M were more frequently observed in the P53-mutated group than in the non-P53-mutated group, although the p value did not reach statistical significance (p ˃ 0.05; Table 9).

Table 6 Comparison of Qualitative Data Obtained on MRI Features Stratified by P53 Status (10mm ≤ LTD ≤ 100mm)

Table 7 Comparison of Qualitative Data Obtained on MRI Features Stratified by P53 Status (10mm ≤ LTD ≤ 20mm)

Table 8 Comparison of Qualitative Data Obtained on MRI Features Stratified by P53 Status (20mm ˂ LTD ≤ 50mm)

Table 9 Comparison of Qualitative Data Obtained on MRI Features Stratified by P53 Status (50mm ˂ LTD ≤ 100mm)

Univariate and Multivariate Analyses

Univariate logistic regression analysis identified six risk factors that were significantly related to P53-mutated HCC, namely, largest tumor diameter, AFP level, the Edmondson–Steiner grade, microvascular invasion, rim enhancement, and corona enhancement. These parameters were analysed using multivariate logistic regression. Edmondson–Steiner grade (odds ratio, 2.280; 95% CI: 1.268, 4.101; p = 0.006) and rim enhancement (odds ratio, 2.517; 95% CI: 1.095, 5.784; p = 0.030) were found to be independent variables associated with P53-mutated HCC (Table 10). When the maximum diameter was greater than or equal to 10mm and less than or equal to 20 mm, univariate and multivariate analyses indicated that one risk factor was significantly associated with P53-mutated HCC, and enhancing capsule was an independent predictor of P53-mutated HCC (odds ratio, 6.200; 95% CI: 1.116, 34.449; p = 0.037; Table 11). When the maximum diameter was between 20 mm and 50 mm, and univariate analysis suggested that Edmondson–Steiner grade, corona enhancement, and nodule-in-nodule architecture were predictors of P53 mutation, and multivariate analysis suggested that corona enhancement (odds ratio, 2.102; 95% CI: 1.022, 4.322; p = 0.043) and nodule-in-nodule architecture (odds ratio, 2.157; 95% CI: 1.033, 4.504; p = 0.041) were independent risk factors for P53-mutated HCC (Table 12). When the maximum diameter was greater than 50 mm and less than or equal to 100 mm, in univariate analysis, the maximum diameter and AFP were associated with P53-mutated HCC, at multivariable regression analysis, increased diameter (odds ratio, 1.035; 95% CI: 1.001, 1.069; p = 0.044) and AFP ≥ 400 (ng/mL) (odds ratio, 3.336; 95% CI: 1.052, 10.577; p = 0.041) were independent variables associated with P53-mutated HCC (Table 13).

Table 10 Univariate and Multivariate Analyses of Risk Factors for the P53-Mutated HCC (10mm ≤ LTD ≤ 100mm)

Table 11 Univariate and Multivariate Analyses of Risk Factors for the P53-Mutated HCC (10mm ≤ LTD ≤ 20mm)

Table 12 Univariate and Multivariate Analyses of Risk Factors for the P53-Mutated HCC (20mm ˂ LTD ≤ 50mm)

Table 13 Univariate and Multivariate Analyses of Risk Factors for the P53-Mutated HCC (50mm ˂ LTD ≤ 100mm)

Discussion

To date, few studies have analysed the sensitivity or specificity of MRI features for diagnosing or predicting the clinical outcome of P53-mutated HCC. In our study, several of the imaging features were related to P53-mutated HCC; P53-mutated HCC was associated with poorer differentiation and was more likely to show peripheral rim-like enhancement on imaging. In patients with small HCC (10 mm ≤ LTD ≤ 20 mm), enhancing capsule was an independent predictor of P53-mutated HCC; in lesions with a diameter ranging from 20 mm to 50 mm, corona enhancement and nodule-in-nodule architecture were more likely to be observed in the images of P53-mutated HCC; and in patients with large HCC (50 mm ˂ LTD ≤ 100 mm), a higher serum AFP level and large tumors may indicate a higher risk of P53-mutated HCC, while MRI features showed little difference.

A recent study showed that HCC with peripheral rim-like enhancement is associated with poor differentiation, a high frequency of macrotrabecular-massive type, and greater susceptibility to microvascular invasion, and peripheral rim-like enhancement is accompanied by increased necrotic areas in the centre of the tumor and frequent expression of the hypoxia-related marker CAIX.¹⁴ Both oncogenes and hypoxia can induce damage to DNA and signal activation in p53.⁴ Based on the findings of previous studies, a high degree of rim enhancement suggests that the tumor is highly malignant and poorly differentiated, which increases the rate of early recurrence after curative resection of HCC, and substantial tumor necrosis leads to a poorer prognosis.^15,16 All of these findings are related to the increased malignancy and excessive growth of tumors, and the acquisition of P53 mutants, a transcription factor, promotes metabolic reprogramming of tumors, causing survival and proliferation of tumor cells.¹⁷ In the present study, poor differentiation and arterial-phase ring enhancement on MRI images were risk factors for P53-mutated HCC, and the results were consistent with these findings.

From 1996 to 2007, the Japanese surveillance system showed a yearly increase in the incidence of small HCCs ≤2 cm,¹⁸ and the treatment modalities used were diverse; therefore, understanding the internal characteristics of small HCC can aid in the development of personalized management. In our study, enhancing capsule was a risk factor for P53-mutated HCC in patients with small HCC. Enhancing capsule of HCC in patients with chronic liver disease is more easily recognized in the late portal venous phase of gadoxetic acid disodium-enhanced MRI.¹⁹ The presence of an enhancing capsule in HCC is caused by slow flow within intracapsular vessels and retention of contrast media in prominent peritumoral sinusoidal spaces or fibrosis.²⁰ Some scholars have shown that the fibrous capsule and the prominent blood sinus overlap, that some microvessels form within the fibrous capsule, that the fibrous capsule accompanies the growth of HCC and that angiogenesis occurs.²¹ Recent studies have reported that dilated vessels in a pattern of vessel-encapsulated tumor clusters, possibly corresponding to neovascularized arteries or dilated sinusoidal gaps, are frequently observed on dynamic CT in P53-mutated HCC,¹² and enhancing capsule is a risk factor for the development of microvascular invasion in HCC.²² Our study demonstrated a close relationship between enhancing capsule and p53-mutated HCC.

Several scholars have studied the preoperative imaging features of solitary HCC (≤5 cm) to identify potential biomarkers that can be appropriately selected to expand the scope of surgical resection and reduce tumor recurrence.²² Corona enhancement is an important ancillary feature of malignancy in LI-RADS and is readily observed in the late arterial and early portal vein phases. The nodule-in-nodule architecture appears mainly because of the morphological proliferation of poorly differentiated cells in the inner nodes replacing the well-differentiated outer nodes. In our study, P53-mutated HCC lesions with diameters ranging from 20 mm to 50 mm exhibited these imaging features because P53 mutation induced these phenotypes. The internal lesions of HCC may accumulate fat,²³ thus leading to T2 isointense or hyperintense nodules, whereas mutant p53 can promote lipid synthesis by altering the activity of various transcription factors or signalling molecules involved in lipid metabolism, such as p63, p73, Nrf2, and AMP-activated protein kinase (AMPK).^24–26 Neovascularization occurs mainly at the periphery of the tumor and is prone to peripheral corona enhancement, which correlates with corona enhancement due to the rapid decrease in central perfusion, which favours central tumor necrosis and may result in the formation of nodule-in-nodule architecture.²⁷ Our study identified both nodule-in-nodule architecture and corona enhancement as risk factors for P53-mutated HCC.

In this study, larger HCC (50 mm ˂ LTD ≤ 100 mm) had diverse imaging features, while tumor diameter and AFP level were risk factors for P53-mutated HCC. Rapidly growing HCC usually exhibits areas of necrosis due to hypoxia;²⁶ thus, larger tumors are prone to cystic degeneration and necrosis, and their imaging features can be affected. Liu et al^28,29 suggesting that increased tumor aggressiveness and infiltrative growth in HCC are associated with larger tumor sizes. P53 is a tumor suppressor gene, but mutant p53 proteins not only lose their tumor suppressor function but also often promote tumorigenesis through various pathways. Larger tumor diameters and AFP levels also negatively affect the survival of HCC patients.³⁰ AFP can be routinely monitored in patients undergoing HCC surveillance because it is an easily available biomarker for HCC. Elevated serum AFP levels indicate a poor prognosis in patients with HCC.³¹ P53 mutations promote hepatocellular carcinoma cell proliferation, and AFP is activated during liver regeneration and hepatocarcinogenesis.³² Therefore, elevated AFP levels are prone to occur in patients with P53-mutated HCC. Vessel-encapsulated tumor clusters (VETC) promote HCC progression and metastasis; for example, one study indicated that 75% of HCC presenting with VETC were larger than 5 cm in diameter, suggesting that this specific vascular pattern is associated with tumor progression,³³ and other studies have demonstrated that CD34-positive vessel-encapsulating tumor clusters are also more likely to be present in P53-mutated HCC.¹² This study also suggested that larger tumor diameters and higher AFP levels are also more likely to be found in P53-mutated HCC patients.

This study has several limitations. First, because this was a single-centre and retrospective study, selection bias was inevitable. Second, we considered that the immunohistochemical expression of P53 indicated a mutation in TP53 (based on the percentage of nuclear staining), which should be confirmed by genomic mapping to improve the accuracy of the results. Third, the tumors in this study were single masses ranging from 10 mm to 100 mm in diameter, and the results cannot be applied to tumors of other diameters or multiple masses. Finally, although preoperative imaging data for P53-mutated HCC patients were obtained, long-term follow-up data are needed to determine the association between outcome and survival.

Conclusion

In conclusion, poor differentiation and rim enhancement on enhanced MRI are useful imaging markers for the diagnosis of P53-mutated HCC. The presence of an enhancing capsule in small HCC (10mm ≤ LTD ≤ 20 mm) can predict P53-mutated HCC, and the presence of corona enhancement and nodule-in-nodule architecture in HCCs with diameters ranging from 20 mm to 50 mm should be highly important for identifying P53-mutated HCC. In large HCC lesions (50 mm ˂ LTD ≤ 100 mm), elevated AFP and enlarged masses should also be emphasized.