Ki-ras Mutations as a Prognostic Factor in Extrahepatic Bile System Cancer By Noria Malats, Miquel Porta, Josep L. Piiiol, Josep M. Corominas, Juli Rife, and Francisco X. Real for the PANK-ras I Project Investigators Purpose: To assess the prevalence and prognostic sig ace of Ki-roscodon 12 mutations in extrahepatic biliary system cancer (EBSC). Patientsand Methods: Patients diagnosed with EBSC between 1980 and 1990 (N = 111) were selected from two hospitals. DNA was amplified from paraffin-embedded tissues and mutations in codon 12 of Ki-ras were detected using the artificial restriction fragment-length polymorphism (RFLP) technique. Results: Tissue was available from 68.5%of patients. The prevalence of mutations was 41%. There was no association between mutations and clinical and pathologic characteristics; however, mutations in Ki-raswere associated with survival, with a median survival duration of 7.7 months for patients with wild-type Ki-ras and 1.7

months for patients with mutated tumors (hazards ratio [HR] = 1.67; P = .075). Among patients with stage I to II tumors, the chance of dying of patients with the mutation was 7.8 times higher than that of patients without the mutation (P = .087); the corresponding HR for patients with stage III to IV disease was 2.9 (P = .003). After adjusting for age, tumor site, histology, differentiation, and stage, the HR for Ki-ras mutations was 2.12 (P

EXTRAHEPATIC

trials, two examples being colon cancer'" and bladder cancer." Regarding EBSC, little is known. Ki-ras mutations have been analyzed in a small number of studies and no detailed reports of their prognostic value are avail-

bile system cancer (EBSC) comprises tumors that arise from the gallbladder, extrahepatic bile ducts, and ampulla of Vater (International Classification of Diseases for Oncology [ICD-O] code 156).' The prognosis of patients with EBSC is poor, except for those with ampullary cancer, which often causes symptoms at early stages. The relatively low incidence of these tumors, presentation at advanced stages, and paucity of therapeutic studies have hampered significant progress in their management.2 Surgical resection is the only curative treatment for EBSC.2 3" Even in patients with resectable tumors, 5-year 4 survival rates are 32% to 61% for gallbladder cancer, . 36 11% to 25% for extrahepatic bile duct cancer, and 34% to 56% for ampullary cancer.78 Adjuvant therapy has not had a major impact on survival in this group of patients and radiation therapy may increase survival.' For patients with more advanced disease, there is no curative therapy, although radiotherapy and chemotherapy may have a palliative effect."' Whereas clinical and pathologic staging, differentiation degree, and tumor location have been found to be associated with prognosis, at present it is not possible to identify patients with resectable tumors who are at higher risk of recurrence." Therefore, the recognition of novel prognostic factors remains a critical goal, particularly in this subgroup. In patients with advanced disease, establishing prognosis more accurately would also be of help in the design of clinical trials. The analysis of the genetic alterations that lead to cancer is beginning to provide new insights into the molecular subclassification of tumors. Molecular alterations are already being used to improve patient selection in clinical

= .026).

Conclusion: Ki-ras codon 12 mutations are an independent prognostic indicator in patients with EBSC. Mutation detection may be of help in the management of these patients. J Clin Oncol 13:1679-1686. © 1995 by American Society of Clinical Oncology.

able. 14-23

The objectives of this study were to assess the prevalence of mutations in codon 12 of the Ki-ras gene in EBSC and to determine their prognostic significance. PATIENTS AND METHODS Clinical and Pathologic Data Patients were identified from the cancer registries at Hospital del Mar (Barcelona) and Hospital Son Dureta (Palma de Mallorca), two

From the Departments of Epidemiology and Immunology, Institut Municipald'Investigacid MWdica and UniversitatAutbnoma de Barcelona; Hospital del Mar, Barcelona; and Hospital Son Dureta, Palma de Mallorca, Spain. Submitted March 1, 1995; accepted March 13, 1995. Supported by Fondo de Investigacidn Sanitaria,Madrid, Spain, grants no. 91/0595 and 92/0007 and by Institut Municipal d'Investigaci6 MWdica, Barcelona, Spain, grant no. 875138/9. Presentedat the V Congreso de la SociedadEspaflola de Oncolgla Mddica, Granda, Spain, April 19-21, 1995; the II Congresso IberoAmericano de Epidemiologia, Salvador de Bahia, Brazil, April 2428, 1995; the XVII Congresso Nacional de la Sociedad Espafola de Anatomia Patoldgica,Barcelona, Spain, April 27-30, 1995; and the 28th Annual Meeting of the Society for Epidemiologic Research, Snowbird, UT, June 21-24, 1995. Address reprint requests to Miquel Porta,MD, Institut Municipal d'Investigacid Midica. UniversitatAutbnoma de Barcelona, Carrer del Dr.Aiguader 80, E-08003 Barcelona., Spain. C 1995 by American Society of Clinical Oncology. 0732-183X/95/1307-0021$3.00/0

Journalof Clinical Oncology, Vol 13, No 7 (July), 1995: pp 1679-1686

Downloaded from ascopubs.org by 37.44.207.162 on January 28, 2017 from 037.044.207.162 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

1679

1680

MALATS ET AL

university hospitals in northeast Spain. All patients with a discharge diagnosis of EBSC (ICD-O 156), registered between 1980 and 1990, were included (N = Ill1). Because ampullary cancer is a distinct clinical entity, separate analyses were conducted where appropriate. Sociodemographic, epidemiologic, clinical, and pathologic data were obtained from clinical and pathologic records through a structured data form. Smoking history (ever/never) and alcohol consumption history (>/< 30 g/d) were recorded as dichotomous variables. Prior medical history of diabetes mellitus, acute and chronic pancreatitis, gallstones, duodenal ulcer, and psychiatric disorders was recorded as present if mentioned in the clinical records, and as absent otherwise. Tumor site was coded according to the ICD-O.' Most valid basis of cancer diagnosis was categorized following cancer 24 registries' recommendations. Tumor extent was classified according to the International Union Against Cancer (UICC) staging system. 25 Tumor differentiation, determined by a single pathologist (J.M.C.). was graded as well, moderate, and poor following ICD-O criteria.' Treatment intention was classified as curative and palliative. A follow-up rate of 94% was achieved and, at the end of the followup period, 25% of subjects were alive. Survival was determined as the period between diagnosis and death/last evaluation. Tissue Specimens An exhaustive search for histologic material from all patients included in the study was performed. Twenty sections of 5-pm thickness were cut from each specimen and placed on glass slides. The first, tenth, and twentieth sections were stained with hematoxylin/ eosin and used for histologic evaluation. Standard precautions were taken to avoid sample contamination. All cases were reviewed by one pathologist (J.M.C.), who was unaware of the original diagnosis or the results of the molecular analysis. Tumor and nontumor areas were defined by microscopic examination. The material corresponding to these areas was consecutively scraped, always starting from the nontumor area. The percentage of tumor cells present in the tumor area was assessed by two independent investigators (J.M.C. and F.X.R.). The tissue surface analyzed ranged from 0.5 to 1 cm'.

Fig 1. PCR-RFLP analysis of Ki-ras codon 12 mutations: test samples (lanes 1-4), molecular size standards (bars, lane 5), negative control (oral mucosal scrapings, lane 6), and positive control (SKPC-1, lane 7). Wild-type sequences are digested (lower arrowhead), whereas mutated sequences are not (upper arrowhead).

tion. To rule out DNA contamination, tubes that contained only water were always included for amplification. The technique was able to detect one homozygous mutated cell in the presence of 10' 2 wild-type Ki-ras cells. 8 The amplification rate was 100%. Figure I shows a representative example of the results of this analysis. A similar artificial RFLP-based technique was used to characterize mutated samples. The details of the methodology have been described elsewhere."7 To detect GGT -* CGT (Gly - Arg) mutations, the product of the second PCR was digested with Mnll. To detect GGT - GAT (Gly -* Asp), GGT -- GT7 (Gly -, Val), and GGT -TGT (Gly -* Cys) mutations, three additional amplification reactions were performed. The products of these reactions were digested with HphI, HincIl, and HindIII, respectively; mutated sequences are cleaved, whereas wild-type sequences are not. Digestion reaction products were analyzed by 12% acrylamide gel electrophoresis and ethidium bromide staining. Interpretation of digestion products' electrophoresis was performed by consensus of two investigators (N.M. and A.S.). When discordant results were obtained, the analysis was repeated and results were evaluated again. The reliability of results' interpretation was further assessed by an independent investigator (F.X.R.). Agreement was high (> 95%) for all enzyme digestions.

Statistical Analysis Detection of Ki-ras Mutations Careful measures were taken to avoid contamination during all steps of amplification and analysis.2 6 Tissue scrapings were deparaffinized in xylene, washed with 95% ethanol, and pelleted. After desiccation and heating at 95 0C for 10 minutes, DNA was purified using a commercial kit (LINUS; Cultek, Madrid, Spain) and amplified using a nested polymerase chain reaction (PCR). To detect Kiras mutations, the artificial restriction fragment-length polymorphism (RFLP) technique was used as described in detail elsewhere,27 with minor modifications. Briefly, 15 cycles of amplification with primers K IUSX and K IDAX were performed. The 225-bp fragment thus obtained was used to amplify a 103-bp internal fragment. The oligonucleotide primers for this reaction (Kl2BstNI and KIDA exon) contain an artificially created endonuclease restriction site. The amplified product was analyzed by acrylamide gel electrophoresis and ethidium bromide staining and an aliquot was digested with BstNI for 18 hours. This enzyme cleaves the 103-bp amplification product that contains wild-type sequences into two fragments of 82 bp and 21 bp, whereas it does not cleave mutated codon 12 sequences. Digestion reaction products were analyzed by 12% acrylamide gel electrophoresis and ethidium bromide staining. DNA from normal cells and from cell lines known to harbor mutations at codon 12 of Ki-ras was included in each amplification and digestion reac-

The comparison of two qualitative variables was performed using Pearson's X2 test or, alternatively, with Fisher's exact test when Ž 20% of cells had expected counts less than five." Odds ratios and their 95% confidence intervals (CIs) were used to estimate the magnitude of associations between variables, and were adjusted by unconditional logistic regression." Student's t test or Mann-Whitney U test was used to analyze the relationship between a categoric variable with two levels and a normally or nonnormally distributed quantitative variable, respectively. Survival curves were computed by the Kaplan-Meier method and compared by the log-rank test." 3 Cox proportional hazards regression was applied to estimate the hazards ratio (HR) of mutation adjusted for potential confounding and effect-modifying variables (tumor site, histology, differentiation degree, and stage). All P values are two-sided. Statistical analysis was performed using the Macintosh statistical packages Statview 4.01 (Abacus Concepts Inc, Berkeley, CA) and SPSS 4.0 (SPSS Inc, Chicago, IL).

RESULTS

The mean age of the 111 patients with EBSC was 69 years (SD, 12 years); 74% were women. A similar proportion of patients came from each hospital: 53% from

Downloaded from ascopubs.org by 37.44.207.162 on January 28, 2017 from 037.044.207.162 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

1681

Ki-rms MUTATIONS IN BILE SYSTEM CANCER Table 1. Median Survival, Log-Rank Test, P Values, and Univariate HRs

Variable Total Sex Female Male Age, years < 61.7 61.7-70.9 71.0-77.0 > 77.0 Hospital Son Dureta del Mar Site Gallbladder Bile duct Ampulla Unknown Histology Adenocarcinoma Other Unknown Differentiation Well Moderate Poor Unknown Stage I II Ill IV Unknown Treatment Curative Palliative Ki-ras analysis Not performed Performed* (a) Wild type Ki-ras Mutated Ki-ras (b) Wild type Ki-ras Mutated Ki-ras

No. of Patients

%

111

Median Survival (months)

P

HR

95% Cl

1.00 0.73

0.44-1.19

1.00 1.36 1.71 1.55

0.74-2.51 0.89-3.27 0.79-3.04

1.00 1.47

0.93-2.33

1.00 1.35 0.42 3.72

0.82-2.22 0.15-1.19 1.62-8.51

1.00 1.37 1.80

0.62-3.01 0.89-3.66

1.00 1.83 3.54 2.34

0.79-4.27 1.51-8.27 1.00-5.51

1.00 2.44 3.00 8.14 7.80

0.76-7.85 0.80-11.27 2.85-23.25 2.27-26.81

1.00 3.73

2.00-6.97

5.92

82 29

73.9 26.1

4.80 11.24

28 28 28

25.2 25.2 24.4 25.2

9.51 5.16 5.15 3.92

52 59

46.8 53.2

10.49 3.81

60 37 7 7

54.1 33.3 6.3 6.3

6.27 5.04 > 48.00 0.88

92 9 10

82.9 8.1 9.0

7.10 5.04 1.00

17 38 31 25

15.3 34.2 27.9 22.6

17.72 9.65 2.39 9.02

12 18 9 64 8

10.8 16.2 8.1 57.7 7.2

> 48.00 25.36 8.57 2.79 0.80

28 83

25.2 74.8

35 76 45 31 42 28

31.5 68.5 59.2 40.8 60.0 40.0

27

Log-Rank

1.603

.205

2.968

.396

2.784

.095

15.953

.001

3.142

.208

10.992

.012

30.903

< .001

31.18 2.91

19.516

< .001

5.84 6.03 7.67 1.67 7.30 1.35

0.114

.736

3.268

.071

8.218

.004

1.00 1.09 1.00 1.67 1.00 2.27

0.67-1.75 0.95-2.93 1.27-4.05

*(a) All patients (n = 7 6); (b) ampullary cancer patients excluded In = 70).

Hospital del Mar and 47% from Hospital Son Dureta. Diabetes mellitus and gallstones were the most common past medical illnesses (20% and 17%, respectively), and only in nine patients was a family member with cancer recorded. Twenty-five percent of patients were smokers and 35% had a history of alcohol consumption. Fifty-four percent of patients (n = 60) had a diagnosis of gallbladder cancer, 33% (n = 37) extrahepatic bile duct cancer, and 6% (n = 7) ampullary cancer. In seven patients, the precise origin of the tumor could not be determined. Histo-

logic confirmation was available for 101 patients (91%) and most of them had adenocarcinomas. Fifty-eight percent of patients (n = 64) presented with a stage IV tumor and only 25% (n = 28) received treatment with curative intention. The distribution of age, sex, hospital, and degree of differentiation of ampullary cancers was not significantly different from the other tumors. Three of the ampullary tumors were stage I and three stage II, while stage was unknown for one. Paraffin-embedded tissue was available from 76 pa-

Downloaded from ascopubs.org by 37.44.207.162 on January 28, 2017 from 037.044.207.162 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

1682

MALATS ET AL

100

100

A

B

All stages 80

s C

p=O.004

60

a,

40 0.

......... . ...

20

r

es.=.

Wild type K-rascases (n=42) Mutated K-ras cases (n=28) 10 0

20

10

30

40

50

60

20

30

40

50

60

70

Months after diagnosis

70

Months after diagnosis 10.

C

Stages Ill-IV

0 002

p= .

ae

Fig 2. Kaplan-Meier survival curves for patients with mutated and wild-type EBSC. (A) All stages; (B) stages I and II; (C) stages III and IV. Ampullary cancer patients (n = 6) were excluded.

C.i

a, 0.

Wild type Ki-ras cases (n=30)

5

25

30

35

Months after diagnosis

tients (75% of 101 patients from whom histologic confirmation had been obtained and 68% of the total 111 patients). According to the main sociodemographic, clinical, and pathologic variables, the group of patients whose tumor tissue was available for Ki-ras mutational analysis was similar to the group for whom it was not. However, patients with extrahepatic bile duct cancer were underrepresented in the former group (26% v 49%, P = .105). The overall prevalence of Ki-ras codon 12 mutations was 41% (31 of 76). According to tumor site, 34% of gallbladder cancers (15 of 44), 50% of extrahepatic bile duct cancers (10 of 20), and 50% of ampullary cancers (three of six) harbored Ki-ras codon 12 mutations. Mutations were also identified among three of six EBSC patients whose tumor subsite was unknown. The proportion of tumor cells in the tissue area analyzed was similar for mutated and wild-type tumors (data not shown). The spectrum of mutations was as follows: 10 patients (32%) had a Val mutation, four (13%) an Asp mutation, two (7%) a Cys mutation, and one a double mutation (Val and Arg), while in 14 patients (45%) the type of mutation was unknown. The characteristics of patients with mutated tumors were compared with those of patients with wild-type sequences. There were no statistically significant associa-

tions with age, sex, hospital source, medical history, pathologic characteristics, or stage. Mutations were more prevalent in metastatic lesions than in primary tumors (68% v 32%, P = .166). Table 1 lists the median survival time for each variable stratum, the log-rank test, and the corresponding univariate HR. A HR greater than 1 indicates that the risk of death is higher than that of patients in the reference category (denoted by a HR of 1.00). The parameters most strongly associated with survival in the univariate analysis were as follows: hospital source (P = .095), tumor site (P = .001), differentiation degree (P = .012), stage (P < .001), and treatment intention (P < .001). Although survival was somewhat better for patients from Hospital Son Dureta, the influence of hospital source on survival disappeared once adjustment was made for these variables (P = .343). Patients with Ki-ras mutated tumors had a shorter median survival time than patients with wild-type tumors (1.67 v 7.67 months, P = .071). Accordingly, the chance of dying of patients with Ki-ras mutated tumors was 1.67 times greater than that of patients without mutations (95% CI, 0.95 to 2.93). When patients with cancer of the ampulla of Vater were excluded, the relative hazard in-

Downloaded from ascopubs.org by 37.44.207.162 on January 28, 2017 from 037.044.207.162 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

1683

Ki-ras MUTATIONS IN BILE SYSTEM CANCER creased to 2.27 (95% CI, 1.27 to 4.05) (Table 1 and Fig 2). Table 2 lists the prevalence of mutations for each stratum of the variables analyzed in Table 1, along with ageadjusted HRs (HRa) for the presence of the mutation, 95% CIs, and corresponding P values. Among women, the chance of dying for patients with mutated tumors was 1.95 times higher than that of patients with wild-type sequences. Similarly, men with mutated tumors had a twofold increase in the risk of death, although the number of patients was too low to achieve statistical significance. Ki-ras mutations were associated with survival in patients from Hospital del Mar (HRa = 2.36, P = .024), patients with extrahepatic bile duct tumors (HRa = 3.32, P = .061), patients with adenocarcinomas (HRa = 1.89, P = .045), patients with poorly differentiated tumors (HRa = 3.94, P = .011), patients who received palliative treatment (HRa = 2.13, P = .019), and those with stage III to IV

Table 2. Stratified Survival Analysis: HRa for Patients With Mutated Versus Wild-Type Tumors Mutated Cases No.

Global Sex Female Male Hospital Son Dureta del Mar Site Gallbladder Bile duct Ampulla Unknown Histology Adenocarcinoma Other Unknown Differentiation Well Moderate Poor Unknown Stage 1-11 I-lit III-IV Unknown Treatment Curative Palliative

%

HRa

95% CI

P

31

40.8

23 8

39.0 47.1

1.95 2.03

0.97-3.93 0.63-6.56

.063 .238

13 18

39.4 41.9

1.25 2.36

0.49-3.22 1.12-4.95

.643 .024

15 10 3 3

34.1 50.0 50.0 50.0

1.87 3.32

0.87-4.03 0.95-11.63 -

.109 .061 -

5.53

0.39-79.32

.208

26 4 1

38.2 66.7 50.0

1.89 0.51 -

1.01-3.54 0.01-37.25 -

.045 .755 -

6 11 13 1

42.9 32.3 50.0 50.0

3.95 1.45 3.94 -

0.55-28.19 0.56-3.74 1.38-11.3

.170 .445 .011 -

8 5 20 3

34.8 29.4 40.0 100

1.33 7.80 2.94 -

0.31-5.73 0.74-81.96 1.45-5.96 -

.700 .087 .003

5 26

22.7 48.1

0.90 2.13

0.19-4.28 1.13-4.02

.892 .019

*All 3 patients with mutations were alive at the end of the follow-up period, while all 3 without mutations were dead. tAmpullary cancer patients excluded.

tumors (HRa = 2.94, P = 0.003). When ampullary cancers were excluded from the analysis, the association between Ki-ras mutation and prognosis was also evident in patients with stage I/II tumors: patients with mutations had a chance of dying 7.8 times greater than those without mutations (P = .087) (Table 2 and Fig 2). Results of Cox's proportional hazards analyses are listed in Table 3. Age was always included in the models because it often affects treatment decisions." When Ki-ras mutation and tumor site were the only variables considered (model 1), a significant risk for mutation was found: the age- and site-adjusted hazard for patients with mutations was almost twice the risk of patients with wild-type tumors (P = .024). Taking gallbladder cancer as the reference category, ampullary tumors had a better prognosis (HR = 0.33, P = .081) and tumors from unknown site had a worse prognosis (HR = 2.77, P = .027), likely as a result of more advanced stage. The inclusion of histology as a covariable (model 2) did not affect the parameters obtained in the first model. When the differentiation degree was included (model 3), the HR for Ki-ras mutation increased to 2.30 (P = .007). In this model, ampullary site lost significance, whereas the mutation-, site- and histology-adjusted risk of death for patients with poorly differentiated tumors was more than five times the risk of subjects whose tumors were well differentiated (P = .002). The exclusion of ampullary cancers did not change substantially the parameters of the previous three models. Finally, when stage was also considered (model 4), only two variables showed a significant HR: stage (P < .01) and mutation (HR = 2.12, P = .026) (Table 3). This result held when only gallbladder and extrahepatic bile duct cancers were considered (model 5); a slight increase in the HR for mutation and a decrease in the HR values for stage are worth noting. These results were not altered when adjustments for sex, hospital, and medical history were made. DISCUSSION The main finding of this study is that Ki-ras mutations are associated with a twofold increase in the risk of death in patients with EBSC. Although the study was retrospective, all patients diagnosed in the two hospitals were identified and patients whose tumors were analyzed were representative of the whole group. Cox's regression analysis showed that the presence of Ki-ras mutations was the only significant independent prognostic factor in addition to disease stage at diagnosis. Tumor site within the extrahepatic biliary system and differentiation degree, two factors that have been reported to have prognostic value in other studies, 31"'4 did not have an independent association with prognosis when Ki-

Downloaded from ascopubs.org by 37.44.207.162 on January 28, 2017 from 037.044.207.162 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

1684

MALATS ET AL Table 3. Multivariate Survival Analysis by Proportional Hazards Modeling Model

Variable Model LRT P Mutation* HR 95% Cl P Site Gallbladdert Bile duct HR 95% CI P Ampulla HR 95% CI P Unknown HR 95% CI P Histology Adenocarcinomat Other HR 95% CI P Differentiation Wellt Moderate HR 95% Cl P Poor HR 95% CI P Stage

It

No. of Patients

1

2

3

4

5

76

13.977 .016

12.994 .043

27.630 < .001

46.600 < .001

39.802 < .001

1.98 1.10-3.58 .024

2.04 1.10-3.78 .025

2.30 1.25-4.24 .007

2.12 1.10-4.11 .026

2.51 1.26-5.01 .009

1.00

1.00

31

44 20

1.00

1.04 0.53-2.04 .919

1.06 0.53-2.14 .869

1.61 0.78-3.33 .196

1.55 0.70-3.45 .283

0.33 0.09-1.15 .081

0.32 0.09-1.14 .079

0.38 0.10-1.42 .152

1.96 0.34-11.37 .452

2.77 1.12-6.86 .027

2.88 1.06-7.82 .039

2.60 0.92-7.34 .072

1.64 0.56-4.83 .368

IV HR

1.41 0.63-3.18 .401

6

68 6

1.00 0.62 0.22-1.73 .361

14 34

1.00 0.40 0.13-1.18 .097

1.00 0.57 0.17-1.88 .355

1.55 0.53-4.56 .426 1.00 0.56 0.17-1.83 .338

1.00

1.00

1.00

1.82 0.67-4.95 .242

0.88 0.28-2.80 .828

0.54 0.16-1.82 .319

5.34 1.82-15.71 .002

1.68 0.50-5.60 .401

1.08 0.31-3.79 .901

1.00

1.00

1.14 0.54-23.62 .186

0.09-14.00 .922

8.27 0.86-79.96 .068

4.21 0.38-47.7 .243

30.70 3.81-247.65 .001

13.86 1.55-123.79 .019

26

3.57

HR 95% Cl P III HR 95% CI P

1.00

6

9 14

II

1.00

6

44

95% CI P NOTE. HRs for the presence of mutation are adjusted by age and by the other variables in the model. Abbreviation: LRT, likelihood ratio test. *Reference category is wild-type tumor. tReference category. Models I to 4 included all patients; in model 5, ampullary cancer patients were excluded.

Downloaded from ascopubs.org by 37.44.207.162 on January 28, 2017 from 037.044.207.162 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

1685

Ki-ras MUTATIONS IN BILE SYSTEM CANCER ras mutations were taken into account. Despite the small number of patients in some subgroups, Ki-ras mutations were associated with worse survival irrespective of tumor location and stage. This finding, which needs to be confirmed in larger, prospective studies, indicates that mutation detection may be of value in the clinical setting for the following reasons: (1) to identify patients with resectable Ki-ras mutated tumors who could benefit from adjuvant treatment, ie, radiotherapy; and (2) to identify patients with unresectable, nonmutated tumors who might benefit from chemotherapy. Furthermore, as operative mortality rates range from 0% to 40%,`2" mutation detection may also contribute to clinical decision-making in patients who are at high surgical risk. Ki-ras mutations have been detected in a variety of human tumors.3" Whereas no relationship with survival has been demonstrated in colon cancer and pancreas cancer, extensive evidence indicates that Ki-ras mutations are associated with an unfavorable prognosis in patients with non-small-cell lung cancer.'-" In contrast, p53 mutations appear to be uniformly associated with poor prognosis in lung," breast,39"4 bladder," and prostate cancer.4 ' The reasons why Ki-ras mutations have a variable biologic significance in different tumor types are currently not known. The prognostic significance of Ki-ras mutations in EBSC was assessed only in one study of 17 patients with ampullary cancers. 2" Although no formal analyses were presented, the investigators reported no correlation between Ki-ras mutations and patient survival. Few studies have analyzed the presence of Ki-ras mutations in EBSC,4-23 and the reported prevalence ranged widely as follows: from 0% to 100% for extrahepatic bile duct cancer, 20 23 from 0% to 54.5% for gallbladder

cancer,14- 16 2 3 and from 0% to 75% for ampullary cancer. 17.20 The prevalences we found in bile duct cancer

(50%) and gallbladder cancer (34%) are similar to those described by Imai et al. 22 The frequency of Ki-ras mutations we found in ampullary cancer (50%) is higher than that described in the three published studies that included more than 10 patients, 's.'"

in which the prevalence aver-

ages 24.5%. A possible explanation is the low number of patients with ampullary cancers we analyzed; nevertheless, diagnostic misclassification due to the difficulties in assuring the real origin of periampullary neoplasm could also partly explain the results.4 2 This is one of the largest series to analyze codon 12

Ki-ras mutations in ESBC; only two of the 10 studies published so far included more patients. However, further work is necessary to overcome the following limitations: (1) patient classification according to site led to a decreased number of patients in each group; (2) mutations at codons 13 and 61, which also have oncogenic potential, were not studied; (3) tissue was not available from all patients diagnosed in the two hospitals during the study period; and (4) the lack of tissue may be associated with more extensive disease, possibly underestimating the prevalence of mutations. Large, prospective studies will enable determination of the clinical usefulness of detecting Ki-ras mutations in patients with EBSC. ACKNOWLEDGMENT We are indebted to A. Serrat and M. Tork for excellent technical assistance, to A. Salas, A. Bulbena, A. Sitges, and colleagues at Institut Municipal d'Investigaci6 M.dica for valuable contributions to the study, and to Puri Barbas, Leo Espaflol, and Helena Martfnez for secretarial support.

APPENDIX PANK-ras I Project investigatorsare asfollows: M. Porta and F.X. Real (Principal Investigators), N. Malats, J.L. Piflol, J.M. Corominas, J. Rifa, M. Andreu, M. Conangla, T. Thomson, M. Galltn, R. Soli, and F. Tous.

REFERENCES 1. Percy C, van Holten V, Muir C (eds): International Classification of Diseases for Oncology (ICD-O). Geneva, Switzerland, World Health Organization, 1990 2. Lotze MT, Flickiger JC, Carr BI: Hepatobiliary neoplasms, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer, Principles and Practice of Oncology (ed 4). Philadelphia, PA, Lippincott, 1993, pp 883-914 3. Henson DE, Albores-Saavedra J, Corle D: Carcinoma of the extrahepatic bile ducts. Histologic types, stage of disease, grade, and survival rates. Cancer 70:1498-1501, 1992 4. Henson DE, Albores-Saavedra J, Corle D: Carcinoma of the gallbladder. Histologic types, stage of disease, grade, and survival rates. Cancer 70:1493-1497. 1992 5. Ouchi K, Suzuki M, Saijo S, et al: Do recent advances in

diagnosis and operative management improve the outcome of gallbladder carcinoma?. Surgery 113:324-329, 1993 6. Tashiro S, Tsuji T, Kanemitsu K, et al: Prolongation of survival for carcinoma at the hepatic duct confluence. Surgery 113:270-278, 1993 7. Monson JR, Donohue JH, McEntee GP, et al: Radical resection for carcinoma of the ampulla of Vater. Arch Surg 126:353-357, 1991 8. Sperti C, Pasquali C, Piccoli A, et al: Radical resection for ampullary carcinoma: Long term results. Br J Surg 81:668-671, 1994 9. Kraybill WG, Lee H, Picus J, et al: Multidisciplinary treatment of biliary tract cancers. J Surg Oncol 455:239-451, 1994 10. Falkson G, MacIntire IM, Moertel CG: Eastern Cooperative Oncology Group experience with chemotherapy for inoperable gallbladder and bile duct cancer. Cancer 54:965-969, 1984

Downloaded from ascopubs.org by 37.44.207.162 on January 28, 2017 from 037.044.207.162 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.

1686

MALATS ET AL

I1. Puntis MCA: Biliary neoplasms. Curr Opin Gastroenterol 7:800-804, 1991 12. Jen J, Kim H, Piantadosi S, et al: Allelic loss of chromosome 18q and prognosis in colorectal cancer. N Engl J Med 331:213-221, 1994 13. Esrig D, Elmajian D, Groshen S, et al: Accumulation of nuclear p53 and tumor progression in bladder cancer. N Engl J Med 331:1259-1264, 1994 14. Almoguera C, Shibata D, Forrester K, et al: Most human carcinomas of the exocrine pancreas contain mutant c-K-ras genes. Cell 53:549-554, 1988 15. Tada M, Omata M, Ohto M: Clinical application of ras gene mutation for diagnosis of pancreatic adenocarcinomas. Gastroenterology 100:233-238, 1991 16. Capelli G, Cronauer-Mitra S, Peinado MA, et al: Frequency and spectrum of mutations at codons 12 and 13 of the c-K-ras gene in human tumors. Environ Health Perspect 93:125-131, 1991 17. Stork P, Loda M, Bosari S, et al: Detection of K-ras mutations in pancreatic and hepatic neoplasms by nonisotopic mismatched polymerase chain reaction. Oncogene 6:857-862, 1991 18. Motojima K, Tsunoda T. Kanematsu T, et al: Distinguishing pancreatic carcinoma from other periampullary carcinomas by analysis of mutations in the Kirsten-ras oncogene. Ann Surg 214:657662, 1991 19. Levi S, Urbano-Ispizua A. Gill R. et al: Multiple K-ras codon 12 mutations in cholangiocarcinomas demonstrated with a sensitive polymerase chain reaction technique. Cancer Res 51:3497-3502, 1991 20. Satoh K, Sawai T, Shimosegawa T, et al: The point mutation of c-Ki-ras at codon 12 in carcinoma of the pancreatic head region and in intraductal mucin-hypersecreting neoplasm of the pancreas. Int J Pancreatol 14:135-143, 1993 21. Scarpa A. Zamboni G. Achille A, et al: ras-Family gene mutations in neoplasia of the ampulla of Vater. Int J Cancer 59:3942, 1994 22. Imai M. Hoshi T, Ogawa K: K-ras codon 12 mutations in biliary tract tumors detected by polymerase chain reaction denaturing gradient gel electrophoresis. Cancer 73:2727-2733, 1994 23. Watanabe M, Asaka M, Tanaka J, et al: Point mutation of Kras gene codon 12 in biliary tract tumors. Gastroenterology 107:1147-1153, 1994 24. MacLennan R. Muir CS, Steinitz R, et al: Cancer Registration and its Techniques, vol 21. Lyon, France, IARC, 1978 25. Spiessl B, Beahrs OH, Hermanek P. et al (eds): TNM Atlas. Illustrated Guide to the TNM/pTNM Classification of Malignant Tumours (ed 3). Berlin, Germany, Springer-Verlag, 1992 26. Orrego C: Organizing a laboratory for PCR work, (ed 3). in

Innis MA GD Sninsky JJ, White TJ (eds): PCR Protocols. A Guide to Methods and Applications. San Diego, CA, Academic, 1990, pp 447-454 27. Berrozpe G, Schaeffer J, Peinado MA, et al: Comparative analysis of mutations in the p53 and K-ras genes in pancreas cancer. Int J Cancer 58:185-191, 1994 28. Urban T, Ricci S, Gange J-D, et al: Detection of c-Ki-ras mutation by PCR/RFLP analysis and diagnosis of pancreatic adenocarcinomas. J Natl Cancer Inst 85:2008-2012, 1993 29. Armitage P, Berry G (eds): Statistical Methods in Medical Research (ed 2). Oxford, England, Blackwell, 1987 30. Kleinbaum DG, Kupper LL, Morgenstern H: Epidemiologic Research. Principles and Quantitative Methods. Belmont, CA, Lifetime Learning, 1982, pp 320-376 31. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 32. Cox DR: Regression models and life-tables. J R Stat Soc [B] 34:187-220, 1972 33. Wanebo HJ, Falkson G, Order SE: Cancer of the hepatobiliary system, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer, Principles and Practice of Oncology (ed 3). Philadelphia, PA, Lippincott, 1989, pp 836-874 34. Alexander F, Rossi RL, O'Bryan M, et al: Biliary carcinoma. A review of 109 cases. Am J Surg 147:503-509, 1984 35. Bos JL: ras Oncogenes in human cancer: A review. Cancer Res 49:4682-4689, 1989 36. Slebos RJC, Kibbelar RE, Dalesio O, et al: K-ras oncogene activation as a prognostic marker in adenocarcinoma of the lung. N Engl I Med 323:561-565, 1990 37. Rosell R, Li S, Skacel Z, et al: Prognostic impact of mutated K-ras gene in surgically resected non-small cell lung cancer patients. Oncogene 8:2407-2412, 1993 38. Mitsudomi T, Oyama T, Kusano T, et al: Mutations of the p53 gene as a predictor of poor prognosis in patients with nonsmall-cell lung carcinoma. J Natl Cancer Inst 85:2023-2929, 1993 39. Allred DC, Clark GM, Elledge R, et al: Association of p53 protein expression with tumor cell proliferation rate and clinical outcome in node-negative breast cancer. J Natl Cancer Inst 85:200206, 1993 40. Thor AD, Yandell DW: Prognostic significance of p53 overexpression in node-negative breast carcinoma: Preliminary studies support cautious optimism. J Natl Cancer Inst 85:176-177, 1993 41. Navone NM, Troncoso P, Pisters LL, et al: p53 protein accumulation and gene mutation in the progression of human prostate carcinoma. J Natl Cancer Inst 85:1657-1669, 1993 42. Porta M, Malats N, Piflol JL, et al: Diagnostic certainty and potential for misclassification in exocrine pancreatic cancer. J Clin Epidemiol 47:1069-1079, 1994

Downloaded from ascopubs.org by 37.44.207.162 on January 28, 2017 from 037.044.207.162 Copyright © 2017 American Society of Clinical Oncology. All rights reserved.