Clinical Outcomes of Metastatic Breast Cancer in Patients Having Both Pseudocirrhosis and Portal Hypertension with Evident Varices

Wei-Li Ma National Taiwan University Hospital Dwan-Ying Chang National Taiwan University Hospital Ching-Hung Lin National Taiwan University Cancer Center Hospital Kao-Lang Liu National Taiwan University Hospital Po-Chin Liang National Taiwan University Hospital Huang-Chun Lien National Taiwan University Hospital Chan-Chuan Hu National Taiwan University Cancer Center Ling-Yun Huang National Taiwan University Hospital Yi-Chun Yeh National Taiwan University Hospital Yen-Shen Lu (  yslu@ntu.edu.tw ) National Taiwan University Hospital

Conclusions: For patients with pseudocirrhosis, clinical pHTN was associated with liver cirrhosis-related complications and shorter median OS. A few patients with pHTN had prolonged OS with effective systemic treatment and aggressive supportive care.
The term "pseudocirrhosis" may have misled researchers and clinicians to overlook the clinical importance of pHTN in affected patients, and investigations of the impact of pseudocirrhosis on survival has been limited. In clinical practice, up to one-quarter of patients have both "pseudocirrhotic" imaging results and concurrent clinically compromised liver function with typical complications associated with pHTN [18]. Therefore, we conducted this retrospective study to investigate how the presence of pHTN in patients with pseudocirrhosis affects the clinical outcomes of metastatic breast cancer.

Patient Selection, Strati cation, And Analytic Data
The medical records from patients with metastatic breast cancer and imaging-diagnosed pseudocirrhosis between 2005 and 2017 were retrospectively analyzed. Among 164 potentially eligible patients, 58 were excluded: 32 had malignancies other than breast cancer, and 26 had incomplete clinical data or imaging ndings. The remaining 106 patients were included in the analysis. Radiologists coded pseudocirrhosis based on the presence of lobular hepatic contours, segmental volume loss, and caudate lobe enlargement (Fig. 1). Patients with advanced breast cancer may have ascites (including malignant ascites), cachexia-related hypoalbuminemia, or splenomegaly without con rmed pHTN; therefore, we used a stringent criterion to de ne clinical pHTN, namely the presence of esophageal or gastric varices on esophagogastroduodenoscopy (EGD). Eligible patients were divided into groups with or without EGDde ned pHTN. Their ages, breast cancer stages, pathology subtypes, and viral hepatitis status upon breast cancer diagnosis were retrieved from medical records. The initial numbers of metastases in liver lobes, other concurrent visceral metastases present at pseudocirrhosis development, and the interval from stage IV disease to documented pseudocirrhosis were also excerpted for analysis. In addition, systemic chemotherapy regimens and the conventional complications of pHTN and decompensated cirrhosis after pseudocirrhosis diagnosis were reviewed, including GI bleeding, thrombocytopenia, hyperbilirubinemia, coagulopathy, and hyperammonemia. Thrombocytopenia was de ned as platelet count of <100,000/μL, hyperbilirubinemia as total bilirubin > 1.5 times the upper limit of normal, coagulopathy as International Normalized Ratio (INR) > 1.5 times the upper limit of normal, and hyperammonemia as blood ammonia > 50 μmol/L. Most patients underwent computed tomography assessments every 3 months as per the Taiwan National Health Insurance regulations. Overall survival (OS) was de ned as the interval from the date of pseudocirrhosis diagnosis until the date of death from any cause.
The National Taiwan University Hospital Research Ethics Committee (NTUHREC) approved this retrospective study and the waived informed consent (NTUHREC No. 201806028RINC). The medical charts were retrospectively reviewed and the all process was carried out in accordance with the regulations of Helsinki.

Statistical Analysis
The categorical variables of patients with and without pHTN were compared using Fisher's exact test. Continuous variables and medians were compared using the Mann-Whitney U test. OS was estimated using the Kaplan-Meier method. Univariate and multivariate Cox proportional hazard models with timedependent covariates were used to explore associations among pHTN, concurrent lung or brain metastases, and other factors with potential prognostic value for determining OS. The p value of < .05 was considered statistically signi cant.

Results
Patient Characteristics Table 1 shows the clinical characteristics of patients with pseudocirrhosis. As revealed by EGD, 32 patients had clinical pHTN and 74 patients were without pHTN. Patient and tumor characteristics between the 2 groups were generally similar. Overall, the median age was 54 years, 73 patients (69%) had recurrent metastatic breast cancer, 66 (62%) had hormone receptor-positive and human epidermal growth factor receptor 2-negative (HR+/HER2−; luminal-type) disease, 32 (30%) had HER2+ disease, and 8 (8%) had triple-negative disease. Serologically, 9 patients were hepatitis B virus surface antigen (HBsAg) positive, and 1 tested positive for anti-hepatitis C virus antibody (anti-HCV). Moreover, 7 of 9 patients positive for HBsAg received antiviral therapy during systemic anticancer treatment. Of the 9 patients positive for HBsAg, 3 had pHTN, whereas the remaining 6, including the only patient who was anti-HCV(+), did not have pHTN. Moreover, 81 (76%) patients had liver metastases in both lobes, and 91 (86%) had 4 or more liver metastases preceding pseudocirrhosis development. In the study population, 43 (41%) and 23 (22%) patients had concurrent lung metastases and brain metastasis, respectively. The median time from the diagnosis of metastatic breast cancer to the development of pseudocirrhosis was 34 months in patients with pHTN and 26 months in those without (p = .315). All patients received considerable pretreatment before pseudocirrhosis was documented (Supplementary Table 1). The majority (86%) of the patients had received at least 4 lines of chemotherapy since metastasis, and 73 (69%) had received 5 or more lines. Among the 66 patients with HR+/HER2− breast cancer, 45 (68%) had received at least 2 lines of endocrine therapy for metastatic disease. Table 2 shows the systemic chemotherapy regimens and complications of pHTN/hepatic decompensation after the diagnosis of pseudocirrhosis. Regardless of the presence of pHTN, the most frequently prescribed drugs were uorouracil (5-FU) and platinum, with these drugs were prescribed for 66% and 47% of patients in the pHTN group as well as 52% and 46% of patients in the non-pHTN group, respectively. Forty-ve patients without pHTN received paclitaxel and docetaxel (61%), compared with 10 patients with pHTN (31%; p = .005). In addition, 21 (28%) patients without pHTN and 3 (9%) patients with pHTN received doxorubicin (p = .042). Patients with pHTN experienced more manifestations implying hepatic decompensation or complications of pHTN. The incidence rate ranged from the lowest of 59% for hyperammonemia to the highest of 91% for thrombocytopenia; 72% of patients experienced GI bleeding, and 84% had received blood transfusion. For patients without pHTN, the most common complication was thrombocytopenia (55%), and only 6 (8%) patients developed GI bleeding. Patients with pHTN had a median OS of only 5 months compared with 13 months in those without pHTN (p = .002; Fig. 2). The median OS of patients with HER2-positive, luminal-type, and triple-negative breast cancer was 16, 9, and 2 months, respectively (p = .001; Fig. 3). Generally, the prognosis of patients with pseudocirrhosis and pHTN is poor; nevertheless, in the present study, with proper treatment and supportive care, 7 (22%) patients had OS exceeding 1 year. Among these patients, 5

Discussion
In this single-institution retrospective analysis, we identi ed 106 metastatic breast cancer patients with pseudocirrhosis on radiology reports within a 13-year interval. To our knowledge, this is the largest series to date; because pseudocirrhosis was diagnosed by a random radiologist, we postulated that these image changes would be conspicuous rather than subtle. In other words, the actuarial patient number may be higher if a systemic image review is performed. Pseudocirrhosis can be complicated by pHTN [13,15,16,18,[26][27][28][29][30][31][32][33]. In contrast to earlier studies that have used imaging criteria to de ne pHTN, we chose a stringent criterion to de ne pHTN: the presence of varices on EGD. Despite the use of this criterion, up to 30% of patients (n = 32) were identi ed as having pHTN, and this too is the largest case series thus far. Oliai et alretrospectively analyzed a consecutive cohort at the University of California Los Angeles and reported that 37 of 199 consecutively treated patients with metastatic breast cancer developed pseudocirrhosis, and of patients with liver metastases, 55% developed pseudocirrhosis [17].
According to that report, pseudocirrhosis in metastatic breast cancers is not rare. However, the term "pseudocirrhosis" is sometimes misleading because some of these patients have hepatic decompensation.
Although our study did not focus on the risk factors and etiologies of pseudocirrhosis, our ndings are consistent with those in a series reported by Oliai et al [17]. In their series, they found that pseudocirrhosis does not occur in the absence of liver metastases, and those who developed pseudocirrhosis had received more lines of systemic treatments; they thus concluded that higher cumulative exposure to systemic treatment may be causative in contrast to the previous assumption that pseudocirrhosis is an adverse effect of a particular therapeutic agent. In our cohort, every patient had preceding liver metastases, 76% had bilateral lobe involvement, and 61% had at least 5 hepatic nodules upon liver metastases. In the entire cohort, 69% had received 5 or more lines of chemotherapies, and among patients who were HR+, 68% had received 2 or more lines of endocrine therapy. The median time from the diagnosis of metastatic disease to the development of pseudocirrhosis was 27 (0-108) months, which is in line with the postulation that pseudocirrhosis usually affects patients with considerable pretreatment and is a late event.
In the literature, limited data on the histology change of pseudocrirrhosis are available. In our study, 1 patient had undergone liver biopsy, and the pathology nding was fatty change and brosis without malignant cells (Fig. 4). The nding was different from those of Young et aland Diamond et al, who have suggested nodular regeneration hyperplasia or intravascular tumor in ltration with extensive stromal brosis and compression of the portal triad vasculature, respectively [13,34]. The pathogenesis may consist of a spectrum and may change along the disease course, which warrants comprehensive research.
Because we de ned pHTN by the presence of esophageal or gastric varices in this study, it was unsurprising to nd that in addition to triple-negative disease, pHTN is also independently associated with poor OS. Patients with pHTN had signi cantly more complications of hepatic decompensation, including hyperbilirubinemia, thrombocytopenia, coagulopathy, and hyperammonemia. Moreover, variceal bleeding can be lethal. We summarize data on 25 breast cancer patients with pseudocirrhosis complicated by esophageal or gastric varices in previous case reports and series in Table  4 [13,15,16,18,[26][27][28][29][30][31][32][33]. Most of these patients had HR+/HER2− disease, and OS was approximately 1 month from the diagnosis of esophageal or gastric varices. In our 32 patients, 23 had GI bleeding episodes, and 11 deaths were associated with variceal bleeding. The median OS of 5 months in this study is congruent with the OS of approximately 1 month reported in previous studies [16,26,30,31]. Poor liver reserve hampered further anticancer treatment. Except for platinum and infusional 5-FU, almost every chemotherapy drug commonly used for breast cancer requires substantial dose modi cation or is even contraindicated in patients with hepatic dysfunction. This explains why cisplatin and infusional 5-FU were most frequently prescribed after pseudocirrhosis was documented in our study. Furthermore, patients with pHTN may have concomitant ascites, nutrition problems, and resultant poor functional performance. Although managing these patients is challenging, 7 patients with pseudocirrhosis and pHTN had OS exceeding 1 year. Moreover, 5 patients had HR+/HER2−disease, and 2 had HER2+ disease. We believe that effective systemic treatment and aggressive supportive care can prolong the survival of such patients.
The limitations of this study include its retrospective design and the heterogeneous patient sample with different breast cancer subtypes and diverse treatment protocols. The study data were recorded between 2005 and 2017; anticancer drugs approved after 2017 might have improved outcomes. This study evaluated the clinical characteristics and outcomes of patients with breast cancer who had pseudocirrhosis with and without pHTN. Additional studies are warranted to investigate other prognostic factors and chemotherapy dosage regimens as well as the detailed pathogenesis of pHTN in patients with pseudocirrhosis.

Conclusion
In our study cohort, pseudocirrhosis mostly developed in patients with luminal-type breast cancer initially diagnosed at an early stage, but these patients had recurrence with multiple liver metastases. Considerable pretreatment (80% had received 5 or more chemotherapy lines) suggests that pseudocirrhosis is a relatively late event. Patients with and without pHTN, as de ned by EGD-documented varices, had signi cantly shorter OS; nevertheless, a few patients survived for longer than 1 year when given aggressive anticancer therapy and intensive supportive care.

Consent for publication
Not applicable.

Availability of data and material
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors have no con icts of interest to declare.       OS of patients with metastatic breast cancer and pseudocirrhosis with and without portal hypertensiona.
The median OS was 5 months in patients with portal hypertension and 13 months in those without (p = .002). OS: overall survival. aEvident esophageal or gastric varices on esophagogastroduodenoscopy.