Multiple Primary Malignancies in Patients With Hepatocellular Carcinoma

A Largest Series With 26-Year Follow-Up

Wei Xu, MD,  Yilei Mao, MD, PhD

Medicine _ Volume 95, Number 17, April 2016北京协和医院肝脏外科毛一雷

Abstract: Multiple primary malignancies (MPMs) are defined as 2 or

more malignancies without subordinate relationship detected in

different organs of an individual patient. Reports addressing MPM

patients with hepatocellular carcinoma (HCC) are rare. We perform a

26-year follow-up study to investigate characteristics and prognosis of

MPM patients associated with HCC due to the scarcity of relative

researches.

We retrospectively analyzed records of 40 patients who were

diagnosed with MPM including HCC at the Departments of Surgery

at Peking Union Medical College Hospital during 1989 to 2010. Their

clinical characteristics and postoperative survival were compared with

those of 448 patients who had HCC only during the study period.

Among the 40 MPM patients, 11 were diagnosed synchronously and

29 metachronously. The most common extra-hepatic malignancies were

lung cancer (15%), colorectal (12.5%), and thyroid carcinoma (12.5%).

MPM patients had a negative hepatitis B virus infection rate (P¼0.013)

and lower median alfa-fetoprotein (AFP) level (P¼0.001). Post-operative

1-, 3-, and 5-year overall survival (OS) rates for MPM patients were

82.5%, 64.5%, and 38.6% respectively, and showed no significant difference

with those of HCC-only patients (84.7%, 54.2%, and 38.3%

P¼0.726). During follow-up, 24 MPM patients died, including 17

(70.8%) who died of HCC-related causes. In univariate analysis, synchronous

diagnosis, higher gamma glutamyltransferase (GGT) and/or

AFP levels, tumor >5 cm and vascular invasion were significantly

associated with shorter OS, but only tumor size was an independent

OS factor in Cox modeling analysis.

HCC should be considered as a potential second primary for all

cancer survivors. Most MPM patients died of HCC-related causes and

showed no significant difference in OS compared with HCC-only

patients. Tumor size of HCC, rather than MPMs itself, was the only

independent OS predictor for the MPM patients.

(Medicine 95(17):e3491)

Abbreviations: AFP = alfa-fetoprotein, ALT = alanine

transaminase, CI = confidence interval, GGT = gamma

glutamyltransferase, HBsAg = hepatitis B virus surface antigen,

HBV = hepatitis B virus, HCC = hepatocellular carcinoma, HCV =

hepatitis C virus, HR = hazard ratio, MPMs = multiple primary

malignancies, OS = overall survival.

INTRODUCTION

Multiple primary malignancies (MPMs) were first

described according to the 1932 definition of Warren

and Gates: each tumor has to present definite attributes of

malignancy, the tumors have to be histological distinctive

and the possibility of one being a metastasis of the other must

be ruled out.1 Thanks to continually improving screening

programs, diagnostic, and treatment methods, survival rates

for newly diagnosed cancer patients are increasing. This

improvement has led to a steady increase in the number of newly

diagnosedMPMpatients.2 In the United States,MPMs constitute

18% of all cancers diagnosed; in European countries, such as the

Czech Republic, the MPM incidence is more than 11%.3

Hepatocellular carcinoma (HCC) ranks fifth in cancer

incidence and third in cancer mortality worldwide.4 Although

less than 1% of MPM patients reported had HCC in 1990s,5

longer overall survival (OS) of oncology patients elevated the

risk of MPM significantly. By 2002, liver cancer was frequently

diagnosed with other major malignant tumors; it was found in

11.5% of all MPMs in Korea.6 MPM patients who develop HCC

over a long-term follow-up are no longer considered unusual,

and clinicians increasingly need to consider the development of

multiple primary cancers with HCC.

Information regarding the MPM patients with HCC is

important, as it could clarify etiological factors and may verify

the need to screen for associated malignancies during patient

follow-up. Understanding of clinicopathological features and

prognostic factors are also needed to facilitate appropriate

management of MPM patients. However, knowledge of characteristic

and outcomes of MPM patients remains limited.

Editor: Zhentian Li.

Received: December 1, 2015; revised: March 16, 2016; accepted: March

31, 2016.

From the Department of Liver Surgery, Peking Union Medical College

Hospital, Chinese Academy of Medical Sciences and Peking Union

Medical College, 1# Shuai-Fu-Yuan, Wang-Fu-Jing, Beijing, 100730,

China.

Correspondence: Yilei Mao, Department of Liver Surgery, Peking Union

Medical College Hospital, Chinese Academy of Medical Sciences and

Peking Union Medical College, 1# Shuai-Fu-Yuan, Wang-Fu-Jing,

Beijing, 100730, China (e-mail: pumch-liver@hotmail.com).

The 9 authors are justifiably credited with authorship, according to the

authorship criteria. In detail: study concept and design: Wei Xu, Yilei

Mao; Acquisition of data:Wei Xu,Wenjun Liao, Penglei Ge, Jinjun Ren,

Haifeng Xu, Huayu Yang; Analysis and interpretation of data: Wei Xu,

Haifeng Xu, Huayu Yang; Drafting of the manuscript: Wei Xu, Yilei

Mao; Critical revision of the manuscript for important intellectual

content: Wei Xu, Xinting Sang, Xin Lu, Yilei Mao; Statistical analysis:

Wei Xu, Haifeng Xu, Huayu Yang; Technical or material support:

Xinting Sang, Xin Lu; Study supervision: Wei Xu, Huayu Yang, Xinting

Sang, Xin Lu; Final approval: Yilei Mao.

This study was supported by National Key Technology Research and

Development Program of China 2012 (grant number BAI06B01),

National Natural Science Foundation of China (81201566), the National

High Technology Research and Development Program (‘‘863’’ Program)

of China (2015AA020303), and the Specialized Research Fund for the

Doctoral Program of Higher Education (20121106110002).

The authors declare no conflict of interest.

Copyright # 2016 Wolters Kluwer Health, Inc. All rights reserved.

This is an open access article distributed under the Creative Commons

Attribution License 4.0, which permits unrestricted use, distribution, and

reproduction in any medium, provided the original work is properly cited.

ISSN: 0025-7974

DOI: 10.1097/MD.0000000000003491

Medicine®

OBSERVATIONAL STUDY

Medicine _ Volume 95, Number 17, April 2016 www.md-journal.com | 1

To our knowledge, only studies with cohorts of 30 patients

or fewer have been performed in Japan or Western countries for

MPM patients with HCC who had received radical hepatectomy.

7,8 The clinicopathologic characteristics and outcomes of

MPM patients are poorly understood, especially in Asian

countries. This retrospective study includes the largest sample

size than any other researches and 26 years follow-up time,

in order to characterize MPM patients and to explore their

long-term prognosis.

METHODS

Between January 1989 and September 2010, 40 patients

with HCC that had been treated with radical hepatectomies were

diagnosed with extra-hepatic primary malignancies at our

institution; we regarded these patients as the target group

(MPM group). Over the same period, 448 others with HCC

only received hepatectomies; these patients were defined as the

control group. In both groups, HCC was diagnosed on the basis

of the histopathology from hepatectomy samples. The extrahepatic

primary malignancies were diagnosed on the basis of

histopathology from resection (36/40) or biopsy (4/40) samples.

A diagnosis of HCC as a second primary malignancy should be

pathologically confirmed, as the liver is a common site for

metastases and imaging findings may be atypical. To avoid the

possibility of misdiagnosis between HCC and metastatic carcinoma,

MPM patients who were diagnosed only by clinical

methods were not included in this analysis. The MPM group

was further classified into synchronous (2 malignancies diagnosed

within a 6-month period) or metachronous (detected more

than 6 months apart). The study protocol was approved by the

Ethics Committee of Peking Union Medical College Hospital.

The preoperative data of patients’ clinical characteristics

including age, sex, family history, serum hepatitis B virus

(HBV), surface antigen (HBsAg), hepatitis C virus (HCV) antibody,

serum alfa-fetoprotein (AFP) were collected, and histopathologic

information regarding tumor number and size, tumor

location, vascular invasion, nodal status, and cirrhotic change in

background liverwere recorded.Tumordifferentiationwas graded

by theEdmondson grading system.9TheTNMstaging systemwas

used to assessHCCstage.Time forHCCsurgeries, blood loss, and

bloodtransfusionwere recorded.Time for surgerieswas definedas

the time fromthe beginning of surgery topatients’ awakeningfrom

anesthesia. Median survival, and cumulative 3-year and 5-year

survival rates were calculated. OS was defined as the interval

between surgery and death or the last date of follow-up. Curative

therapy for the extra-hepatic primarymalignancieswas defined as

treatmentwithintent tocure, suchas the surgeries formalignancies

of breast, thyroid, digestive system, and respiratory system, or

radial or chemical therapy formalignancies of blood system,while

other treatment methods were regarded as palliative therapy.

Clinical and pathological factors were compared using

either Fisher exact test or Pearson x2-test, as appropriate. The

survival rate was calculated using the Kaplan–Meier method.

COX-regression analysis was performed to identify independent

risk factors with hazard ratio (HR) and 95% confidence interval

(CI). P<< span="">0.05 was considered statistically significant. Data

analysis was performed using SPSS 19.0 software.

RESULTS

Patient Characteristics

Of the 40 MPMpatients, 11 were diagnosed synchronously,

and 29 metachronously, with HCC; 18 patients’ extra-hepatic

primary malignancies occurred prior to their HCC diagnoses

(prior group), and 11 after their HCC diagnoses (post group). The

most sites preceding or following HCC diagnoses were lung (6/

40, 15%), colorectal (5/40, 12.5%), thyroid (5/40, 12.5%), breast

(3/40, 7.5%), prostate (3/40, 7.5%), and sensory organs (3/40,

7.5%); 26 patients were treated by curative therapy and 4 by

palliative therapy (Table 1).

Although diagnostic intervals between the 2 cancers ranged

from 10 months to 21 years in the metachronous group

(68.17_73.99 months), 51.7% (15/29) of the metachronous

patients were diagnosed with secondary cancers within 3 years

of the initial cancer diagnosis (Figure 1). Moreover, 27.6% (8/

29) of theMPMgroup were diagnosed after more than 6 years—

all in the prior group, whose median interval time was significantly

longer than that of the post group (93.89_84.26 months

vs. 26.09_10.98 months, P¼0.003).

The MPM group included 36 men and 4 women. We

detected HBsAg in 57.5% (23/40) patients; HCV antibody

was positive in 17.5% (7/40); cirrhosis was present in 62.5%

(25/40). Interestingly, we found that the proportion of patients

with larger tumors (diameter >5 cm) in the synchronous group

was significantly higher than that in metachronous group (9/11

vs.12/29 P¼0.034; Table 2). No other significant differences

were found between the synchronous and metachronous groups.

Compared clinicopathological features between MPM

patients and HCC patients in control group are shown in

Table 3. The mean age of diagnosis in the MPM group was

significantly older than that in the control group (62.58_11.32

TABLE 1. Site Distribution of Extra-Hepatic Primary Malignancies

in Patients With HCC

Metachronous

Group

Location

Synchronous

Group

Prior

Group

Post

Group Total

Digestive system 4 (10%) 4 (10%) 4 (10%) 12 (30%)

Esophagus 0 1C 1C 2

Stomach 0 2C 0 2

Small intestine 1C 0 0 1

Colorectal 2C 1C 2C 5

Gall bladder 1C 0 1C 2

Head and neck 2 (5%) 4 (10%) 4 (10%) 10 (25%)

Thyroid 0 3C 2C 5

Sensory organ 1C 0 2C 3

Vocal cord 1C 1C 0 2

Respiratory

system

2 (5%) 2 (5%) 2 (5%) 6 (15%)

Lung 2C 2C 21C1P 6

Urinary system 2 (5%) 2 (5%) 1 (2.5%) 5 (12.5%)

Prostate 0 2P 1P 3

Ureter 2C 0 0 2

Breast 0 3C (7.5%) 0 3 (7.5%)

Skin 0 2C (5%) 0 2 (5%)

Nervous system 0 1C (2.5%) 0 1 (2.5%)

Blood system 1C (2.5%) 0 0 1 (2.5%)

Total 11 18 11 40

Treatments for extra-hepatic primary malignancies include curative

therapy (C) and palliative therapy (P). HCC¼hepatocellular carcinoma.

Xu et al Medicine _ Volume 95, Number 17, April 2016

2 | www.md-journal.comCopyright # 2016 Wolters Kluwer Health, Inc. All rights reserved.

years vs. 55.69_11.73 years, P<< span="">0.001). Although more than

half of MPM patients’ HBsAg statuses were positive (57.5%),

the proportion of patients in positive HBsAg status in control

group was significantly higher (76.3%) (P¼0.013). Further,

more patients in control group showed abnormal serum AFP

level (P¼0.001). However, no pathological features showed

significant differences between the 2 groups.

Surgical Procedures

All MPM patients underwent surgeries for HCC including

radical liver resections, as bi- or double segmentomies (n¼19),

single segmentomies (n¼7), left lateral sectorectomies (n¼4),

right anterior sector-plus segmentomies (n¼3), right anterior

sectorectomies (n¼2), right hepatectomies (n¼2), right

posterior sectorectomy (n¼1), left hepatectomy (n¼1), and

left hepatectomy plus segmentomy (n¼1). Simultaneously, 4

patients underwent removal of portal vein tumor thrombi, 4 had

extra-hepatic primary malignancies resected; 2 underwent cardiac

peripheral vascular disconnections; and 3 received lymph

node dissections because of enlarged nodes in the hepatoduodenal

ligament region, including 2 found by intraoperative

exploration and 1 whose suspected lymph node metastasis

was diagnosed by preoperative magnetic resonance imaging

(MRI). In 17 patients, we used Pringle’s maneuver for intermittent

hepatic inflow occlusion during surgery. Median

surgery time was 180 min (range: 100–420 min) and median

blood loss was 225mL (range: 100–2000 mL). No patients died

in the perioperative period.

Surgeries for the HCC-only patients included radical liver

resections as bi- or double segmentomies (n¼143), single segmentomies

(n¼93), right anterior sectorectomies (n¼59), right

posterior sectorectomies (n¼42), left lateral sectorectomies

(n¼25), right hepatectomies (n¼24), left hepatectomies

(n¼24), right anterior sector-plus segmentomies (n¼16), left

half liver sector-plus segmentomies (n¼9), right half liver

sector-plus segmentomies (n¼8), and right half liver plus left

lateral sectorectomies (n¼5). Fourteen patients underwent

removal of portal vein or inferior vena cava tumor thrombus,

FIGURE 1. Diagnosis of secondary cancer by follow-up time after

diagnosis of the first primary tumor, among patients whose first

cancers were HCC (post), whose secondary cancers were HCC

(prior), and those whose cancers were discovered more than 6

months apart (metachronous).

_

The post and prior groups differed

significantly at interval time >72 months (P<< span="">0.05). HCC ¼

hepatocellular carcinoma.

TABLE 2. Comparison of Clinicopathological Characteristics

Between Patients With Synchronous Group and Metachronous

Diagnoses

Characteristics

Synchronous

Group

Metachronous

Group P_

Age, yy 0.173

_62.5 (n¼18) 7 11

>62.5 (n¼22) 4 18

Sex 0.560

Male (n¼36) 11 25

Female (n¼4) 0 4

HBsAg status 0.079

Negative (n¼17) 2 15

Positive (n¼23) 9 14

HCV antibody 0.159

Negative (n¼33) 11 22

Positive (n¼7) 0 7

ALT, U/L 1.000

_40 (n¼20) 6 14

>40 (n¼20) 5 15

GGT, U/L 0.715

_67 (n¼26) 8 18

>67 (n¼14) 3 11

AFP, ng/mL 0.147

_20 (n¼26) 5 21

>20 (n¼14) 6 8

CA 19–9, U/mL 0.182

_37 (n¼32) 7 25

>37 (n¼8) 4 4

Tumor size, cm 0.034

_5 (n¼19) 2 17

>5 (n¼21) 9 12

Tumor location 1.000

Right liver (n¼29) 8 21

Left liver (n¼7) 2 5

Both (n¼4) 1 3

Multiple tumors 0.319

No (n¼34) 8 26

Yes (n¼6) 3 3

Vascular invasion 0.298

No (n¼35) 11 24

Yes (n¼5) 0 5

Edmondson grade 0.728

I–II (n¼24) 6 18

III–IV (n¼16) 5 11

Cirrhosis 0.158

No (n¼15) 2 13

Yes (n¼25) 9 16

Nodal status 1.000

Negative (n¼39) 11 28

Positive (n¼1) 0 1

TNM staging 1.000

I–II (n¼36) 10 26

III–IV (n¼4) 1 3

‘‘Bold’’ value means the ‘‘P’’ value is less than 0.05.

AFP¼alfa-fetoprotein, ALT¼alanine transaminase, GGT¼gamma

glutamyltransferase, HBsAg¼hepatitis B virus surface antigen,

HCV¼hepatitis C virus.

_

Fisher exact test or Pearson x2-test.

yPatients’ age was divided by the median age.

Medicine _ Volume 95, Number 17, April 2016 Multiple Primary Malignancies Associated With Hepatocellular Carcinoma

Copyright # 2016 Wolters Kluwer Health, Inc. All rights reserved. www.md-journal.com | 3

6 patients underwent splenectomy and cardiac peripheral vascular

disconnection, 2 patients underwent phemister surgery

simultaneously, and 200 patients underwent inflow vascular

occlusion using Pringle’s maneuver as mentioned above. Median

time for surgery was 200 min (range: 60–600 min) and median

blood loss was 400mL (range: 50–15,000 mL). Four patients

died in the perioperative period.The 2 groups did not significantly

differ in surgery time (P¼0.099) or blood loss (P¼0.130).

Patient Prognosis

Median follow-up time after HCC surgeries was 41.5

months (range: 2 months to 8.2 years). During the follow-up,

13 (32.5%) patients were still alive, 17 (42.5%) patients died of

HCC-related causes, 2 (5%) of extra-hepatic primary malignancies-

related causes and 5 (12.5%) of unclear causes. Three

(7.5%) patients were unconnected for various reasons. Postoperative

1-, 3-, and 5-year survival rates for the 40 MPM

patients were 82.5%, 64.5%, and 38.6%, respectively.

The effects of clinicopathological characteristics on survival

were evaluated. Synchronous diagnosis, higher levels of

GGT and AFP, tumor diameter >5 cm, and vascular invasion

were significantly associated with poorer OS in univariate

analysis (Table 4), but in Cox-multivariate analysis, only tumor

size remained an independent predictor of survival (Table 5).

The impact of second primary tumor on HCC survival was

also estimated. Post-operative 1-, 3-, and 5-year survival rates

for 448 HCC-only patients were 84.7%, 54.2%, and 38.3%,

respectively, and did not significantly differ from those of the

MPM group (P¼0.726, Figure 2C).

DISCUSSION

Patients with malignancies have received increasing survival

benefits from continuous progress in early cancer detection,

diagnostic sub-classification, and targeted treatments.

Along with increased life expectancy, cancer survivors are at

higher risk of developing another malignancy compared with

the general population. Reportedly, the prevalence of MPMs

has increased, and 11.0% to 21.0% of all cancers have more than

one primary in Western countries.10 The Surveillance, Epidemiology

and End Results Program of the US National Cancer

Institute estimated that 7.9% of cancer survivors were living

with a history of more than 1 primary malignancy and MPMs

now account for 16% of the newly diagnosed malignancies.11

Further, any survivor of cancer has twice the probability of

developing a new second primary cancer than a cancer-free

individual of the same age and sex.12 Thus, an increasing need

exists to determine subsequent cancer risks, and to provide

appropriate surveillance and management. Case reports or

small-sized studies of MPMs that include HCC have been

published in recent years,13–15 but information about their

characteristics and outcomes is still limited, especially for those

who underwent surgeries for HCC. In our series we had 40

MPM patients, the largest sample size ever, receiving radical

resections for HCC and were diagnosed basis on their histopathology.

Although the etiology of HCC in MPM patients remains

unclear, some evidence may be provided by their clinical

features. HCC commonly arises in a background of chronic

hepatitis and cirrhosis in Asian countries.16 In our study, 57.5%

MPM patients had positive HBsAg statuses, which was significantly

less than that the HCC-only control group (76.3%). HCV

infection has also been suggested as a potential risk factor for

HCC. Our study showed that 17.5% patients in MPM group

TABLE 3. Compared Clinicopathological Characteristics

Between MPM Group and Control Group

Characteristic

MPM Group

(n¼40)

Control Group

(n¼448) P_

Age, y 62.58_11.32 55.69_11.73 <0.001< span="">

Sex 0.275

Male 36 369

Female 4 79

Family history

of malignancies

0.516

No 31 371

Yes 9 77

HBsAg status 0.013

Negative 17 106

Positive 23 342

HCV antibody 0.072

Negative 33 412

Positive 7 36

ALT, U/L 0.406

_40 20 258

>40 20 190

GGT, U/Ly 0.317

_67 26 235

>67 14 187

AFP, ng/mLy 0.001

_20 26 161

>20 14 260

Tumor size, cm 0.243

_5 19 260

>5 21 188

Tumor location 0.729

Right liver 29 297

Left liver 7 85

Both 4 66

Multiple tumors 0.830

No 34 366

Yes 6 82

Vascular invasion 0.820

No 35 384

Yes 5 64

Edmondson grade 0.606

I–II 24 290

III–IV 16 158

Cirrhosis 0.283

No 15 131

Yes 25 317

Nodal status 0.403

Negative 39 443

Positive 1 5

TNM staging 0.124

I–II 36 428

III–IV 4 20

‘‘Bold’’ value means the ‘‘P’’ value is less than 0.05.

AFP¼alfa-fetoprotein, ALT¼alanine transaminase, GGT¼gamma

glutamyltransferase, HBsAg¼hepatitis B virus surface antigen,

HCV¼hepatitis C virus, MPM¼multiple primary malignancies.

_

Fisher’s exact test or Pearson’s x2-test.

y1.34% and 6.25% data in control group was missing.

Xu et al Medicine _ Volume 95, Number 17, April 2016

4 | www.md-journal.comCopyright # 2016 Wolters Kluwer Health, Inc. All rights reserved.

were HCVt and did not significantly differ from the HCC-only

group. We are not surprised at the difference in HBsAg infection

between the 2 groups, as reasons for HCC development in

MPM patents may be more complex than those for the HCConly

group, although HBV and HCV infections were regarded

as major causes for HCC.

MPM has been attributed to iatrogenic, environmental, and

hereditary factors.17 Iatrogenic factors, such as anticancer treatments

or radiation therapy, were considered as causes of MPM

tumors. Reportedly, about 40% of patients with metachronous

MPM had histories of receiving anticancer treatments or radiation

therapy to attempt to cure their first cancers and consequently

developed secondary tumors following their initial

treatment.18 In our series, 37.9% (11/29) of MPM patients in the

metachronous group had received chemotherapy or radiotherapy.

Although we might have further considered the effects of

radiation or chemical regimens, age at radiation exposure, and

subsequent treatments, no clear differences were observed

because of insufficient information.

Hereditary factors may be another cause of MPM tumors.

Family history of malignancies, which is regarded as a risk

factor for HCC, may also portend HCC development as a

second malignancy. In our MPM group, 22.5% (9/40) had

immediate family members with histories of cancer, which

was similar to the patients in the HCC-only group (17.2%

77/371). We hypothesize that hereditary factors play a role

in the process, but not solely in MPMs.

Aging is an important etiological factor in MPM patients.

Using the Osaka Cancer Registry data, Tabuchi et al19 reported

that 10-year cumulative risk of metachronous second primary

cancer in Japanese male patients was 10.2% at 50 to 59 years of

age, 16.2% at 60 to 69 years of age, and 21.8% at 70 to 79 years

of age. In the present study, the mean age of HCC diagnosis

in MPM patients was 62.58_11.32 years, which was significantly

older than that of the HCC-only control (55.69_11.73

years). Furthermore, the mean ages of diagnosis did not significantly

differ between the synchronous and metachronous

groups (60.18_8.86 vs. 63.48_12.14 years, P¼0.418). This

implies that older people have higher risks of developing

second malignancies, without the choice for synchronous or

metachronous.

Other risk factors such as BMI, immune status, and

behavior change after the first primary malignancy may also

contribute to HCC development in MPM patients,20 but more

detailed investigation is needed. In most cases, inherited,

iatrogenic, or viral factors are implicated; in other cases a clear

etiopathogenesis is difficult to find, especially for synchronous

MPMs. In our study, 2 synchronous HCC lesions without

cirrhosis in background liver were surprisingly diagnosed

by pathology after surgery for what were thought to be liver

metastasis. One extra-hepatic synchronous tumor was unexpectedly

found during the surgery, which was regarded

as a benign lesion. Thus, the mechanism still needs further

clarification.

TABLE 4. Univariate Analysis of Survival Risk Factors for MPM

Patients

Characteristic

3-Year

Survival

rate (%)

5-Year

Survival

rate (%) P

Age, y_ 0.106

_62.5 (n¼18) 50.0 25.0

>62.5 (n¼22) 72.7 48.0

Gender 0.550

Male (n¼36) 61.0 35.9

Female (n¼4) 75.0 50.0

HBsAg status 0.404

Negative (n¼17) 70.1 44.6

Positive (n¼23) 56.5 31.4

HCV-Ab 0.448

Negative (n¼33) 63.5 39.4

Positive (n¼7) 57.1 28.6

ALT, U/L 0.699

_40 (n¼20) 64.6 26.4

>40 (n¼20) 60.0 45.0

GGT, U/L 0.009

_67 (n¼26) 72.9 47.5

>67 (n¼14) 42.9 17.1

AFP, ng/mL 0.023

_20 (n¼26) 76.7 43.3

>20 (n¼14) 35.7 26.8

Cirrhosis 0.280

No (n¼15) 80.0 43.6

Yes (n¼25) 56.0 34.5

Tumor size, cm < 0.001

_5 (n¼19) 89.5 67.7

>5 (n¼21) 37.5 6.3

Multiple tumors 0.727

No (n¼34) 61.6 38.1

Yes (n¼6) 66.7 33.3

Vascular invasion 0.021

No (n¼35) 62.7 43.2

Yes (n¼5) 60.0 0

Edmondson grade 0.839

I–II (n¼24) 58.8 39.2

III–IV (n¼16) 68.8 34.4

Synchronous or metachronous 0.044

Synchronous (n¼11) 45.5 13.6

Metachronous (n¼29) 68.8 45.9

Treatments for extra-hepatic

primary malignancies

0.614

Curative therapy (n¼26) 67.6 49.4

Palliative therapy (n¼4) 50.0 50.0

‘‘Bold’’ value means the ‘‘P’’ value is less than 0.05.

AFP¼alfa-fetoprotein, ALT¼alanine transaminase, GGT¼gamma

glutamyltransferase, HBsAg¼hepatitis B virus surface antigen,

HCV¼hepatitis C virus, MPM¼multiple primary malignancies.

_

Patients’ age was divided by the median age.

TABLE 5. Cox Analysis of Survival Risk Factors for MPM

Patients

Characteristic P HR (95% CI)

GGT level 0.501 1.459 (0.486–4.377)

AFP level 0.459 1.409 (0.527–3.772)

Tumor size <0.001 1.455 (1.184–1.788)

Vascular invasion 0.065 3.504 (0.927–13.243)

Synchronous or Metachronous 0.111 0.438 (0.159–1.207)

‘‘Bold’’ value means the ‘‘P’’ value is less than 0.05.

AFP ? alfa-fetoprotein, GGT ? gamma glutamyltransferase.

Medicine _ Volume 95, Number 17, April 2016 Multiple Primary Malignancies Associated With Hepatocellular Carcinoma

Copyright # 2016 Wolters Kluwer Health, Inc. All rights reserved. www.md-journal.com | 5

Information about common sites of extra-hepatic malignancies

may improve early detection in high-risk individuals.21

Gastric cancer has been reported as the most common extrahepatic

malignancy among MPM patients with HCC by

Takayasu et al,22 along with colorectal cancer by Ferna´ndez-

Ruiz et al,23 and nasopharynx cancer by Zeng et al.24 Unlike

these previous findings, our study showed that the most common

extra-hepatic malignancy was lung, followed by colorectal

and thyroid. This circumstance may be partly attributable to

different regions from which the study subjects were selected,

for the most common forms of extra-hepatic malignancies were

similar to the most common tumor types in China;25–28 and

partly to the wide variation of multiple cancer distribution,

which may occur as a result of random chance. Contrary to our

expectation, screening for other possible malignancies in cancer

survivors based on the most common sites is difficult because of

the variable distribution of the extra-hepatic malignancy and

any enrichment patterns can hardly have been proven by

statistics yet. Establishment of a pair-wise association with

HCC requires a more systematic and controlled approach.

Previous studies indicate that patients who initially presented

with thyroid, urinary bladder, prostate, cervical, and

uterine cancers were more liable to develop second malignancies,

whereas those with hepatic cancers rarely developed a

second malignancy. They hypothesized that this was, as HCC

has a poor prognosis, HCC patients did not survive long enough

to develop second primaries.29 HCC was among the four cancer

sites with the lowest survival rates and consequently, the shortest

duration of follow-up.30 However, about 40% (11/29) of our

metachronously diagnosed patients were in the post group. The

poor prognosis of HCC patients apparently does not affect the

incidence of another primary tumor occurrence, and the possibility

of developing extra-hepatic malignancies in HCC patients

should not be ignored. Only the obviously longer interval time

of the prior group can be explained partly by poor OS for HCC.

Our MPM patients with interval times longer than 72 months

were all in the prior group (Figure 1). In view of this pattern,

physicians must consider the onset of HCC for each neoplasm,

even many years after first diagnosis.

No consensus currently exists for a method of calculating

the survival rate of MPM patients. Earlier researchers recommended

basing the rate from the diagnosis of the final malignancy

tumor, while others suggest calculating survival from the

diagnosis of the first tumor, to account for the increased risk of

malignancy during the first survival period.31 We focused on

survival time after surgeries for HCC because most MPM

patients died of HCC-related causes, which may indicate that

MPM prognosis is largely determined by survival time after the

HCC surgery. However, this may avoid the bias brought by

longer intervals between MPM diagnoses, which could indicate

a longer survival time.

Survival of MPM patients is reportedly similar to that of

patients with single primary tumors.23,32,33 We had the same

findings for post-surgical survival time (Figure 2C). Further, we

found no significant difference in surgery-related parameters,

such as surgery time and amount of bleeding, between MPM

and control group. We speculate that a history of extra-hepatic

tumor is not a direct obstacle to HCC resection. MPMitself does

not necessarily indicate a poor prognosis, as long as adequate

diagnosis and management are performed. However, HCCrelated

causes predominantly lead to MPM patients’ deaths;

only 38.6% of patients in this study were still alive 5 years after

their liver surgeries.

Male sex and old age have been shown by several studies to

be risk factors for shorter survival in MPM.34 However, we

found no statistical difference for OS in these terms. In the

present study, serum GGT level, AFP level, tumor size, vascular

invasion, and synchronous or metachronous diagnosis led to

distinct outcomes. We verified these results with a Cox multivariate

model, which only found tumor size, as a pathological

feature, to be a significant independent risk factor for survival.

This is an important new observation for patients who survived

their first primary malignancy. Early detection and surgery for

HCC would help improve OS in these patients.

Although the metachronous and synchronous groups significantly

differed in OS (Figure 2A), metachronous or synchronous

diagnoses were not independent OS factors in

multivariate analysis. Metachronous malignant lesions were

discovered because of careful follow-up of the first malignancy,

during which extensive surveillance is carried out to locate

possible metastases. This may explain why HCC lesions in

metachronous group (mainly in the prior group) were found as

smaller tumors than in the synchronous group, which may offer

longer survival. Moreover, no significant difference in OS was

found between the prior and post groups, which demonstrate

that whether extra-hepatic malignancy was the initial or secondary

malignancy did not influence OS after surgeries for

HCC (Figure 2B). Another hypothesis is that as more time

elapses between the 2 primary malignancies, the better the

prognosis. Although the post group has a longer median period

than the synchronous group before diagnosis of second malignancies,

their OS rates did not significantly differ (P¼0.239).

We found no relationship between second primary tumor

development and MPM survival rate of MPM patients.

FIGURE 2. Comparisons of Kaplan–Meier curves between synchronous and metachronous groups (A); prior and post groups (B), and

MPM and control groups (C). MPM ? multiple primary malignancies.

6 | www.md-journal.comCopyright # 2016 Wolters Kluwer Health, Inc. All rights reserved.

In summary, MPMs associated with HCC is rare. Our study

provides the largest sample size of MPM patients ever, receiving

radical resections for HCC. MPM patients were more likely

to die of HCC-related causes even after receiving radical

resection for HCC. Tumor size, rather than MPM itself, was

the only independent predictive factor for OS in MPM patients.

Follow-up for patients recovering from a first malignancy must

be strictly observed, which could improve their chances for

long-term survival. Because of the complex etiology and the

variety ofMPMcancer distributions, HCC should be considered

as a potential second primary for every cancer survivors, even if

not infected by HBV. Additionally, HCC patients, especially

elderly ones, all malignancies must be considered risks of

second tumor.

This study is subject to the limitations inherent in retrospective

work with observation data collected at the specific

point. It also represents the experience of a single tertiary referral

center, and might not be generalized. The etiology of MPMs

remains unclear, because risk factors known to be important to

etiology, such as the details of chemotherapy or radiation therapy,

could hardly be estimated in this study. Limitations of our study

also include the confined sample size, although we have the

largest sample size.Alarger,multi-center study of patients froma

multi-geographic patient base would be more conclusive.