Ann Surg Oncol (2013) 20:233–241 DOI 10.1245/s10434-012-2592-z

ORIGINAL ARTICLE – PANCREATIC TUMORS

Prognostic Relevance of Number and Ratio of Metastatic Lymph Nodes in Resected Pancreatic, Ampullary, and Distal Bile Duct Carcinomas Ewa Pomianowska, MD1,2, Arne Westgaard, PhD1,3, Øystein Mathisen, PhD2, Ole Petter F. Clausen, PhD4, and Ivar P. Gladhaug, PhD1,2 Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo, Norway; 2Department of Hepato-pancreatobiliary Surgery, Oslo University Hospital, Rikshospitalet, Oslo, Norway; 3Department of Oncology, Oslo University Hospital, Norwegian Radium Hospital, Oslo, Norway; 4Department of Pathology, Oslo University Hospital, Rikshospitalet, Oslo, Norway

1

ABSTRACT Background. Lymph node ratio (LNR) may be more useful than nodal (N) status in prognostic subclassification of adenocarcinomas after pancreatoduodenectomy. Ampullary (AC), biliary (DBC), and pancreatic (PC) adenocarcinomas are biologically distinct, and nodal involvement may have different prognostic importance among these separate cancers. Methods. We included 179 consecutive pancreatoduodenectomies for PC, AC, or DBC, and performed standardized histopathologic evaluation, including prospective registration and retrospective reevaluation of the cancer origin. Associations between histopathologic variables and LNR, N status, and number of metastatic nodes were evaluated. Unadjusted and adjusted survival analysis was performed. Results. Overall 5 year survival was 6 % for PC (n = 72), 26 % for DBC (n = 46), and 46 % for AC (n = 61). Lymph node involvement was more frequent in PC (75 %) than in AC (48 %) and DBC (57 %). In PC, N status did not discriminate between prognostic groups (N1 vs. N0; p = 0.31). However, increasing LNR was associated with poorer survival in unadjusted analysis, as well as when adjusting for margin involvement, degree of Electronic Supplementary Material The online version of this article (doi:10.1245/s10434-012-2592-z) contains supplementary material, which is available to authorized users. Ó Society of Surgical Oncology 2012 First Received: 23 May 2012; Published Online: 15 August 2012 E. Pomianowska, MD e-mail: [email protected]

differentiation, and tumor diameter (p = 0.032; hazard ratio 1.87, 95 % confidence interval 1.06–3.31). In AC and DBC, N status clearly discriminated between subgroups of patients with different long-term survival in unadjusted and adjusted survival analysis (N1 vs. N0; p \ 0.001), whereas number of metastatic nodes and LNR did not predict survival among node-positive resections. Conclusions. The predictive value of nodal involvement depends on the type of cancer within the pancreatic head. In AC and DBC, N status adequately discriminates between good and poor prognosis. In PC, LNR may be more powerful in prognostic subclassification.

Primary adenocarcinomas located in the pancreatic head may arise from the ampulla, the periampullary duodenum, the distal bile duct, or the pancreatic tissue itself. Because of the topological proximity in the same anatomic compartment for these structures, the same procedure, i.e., curative-intent pancreatoduodenectomy, is recommended for resectable adenocarcinomas arising from any of these four locations. The precise tumor origin may be difficult to determine, even after standardized histopathological examination of the resected specimen.1 The considerable variation in reported frequencies for the individual tumor sites suggests that the applied method for histopathological determination of the cancer origin varies widely among institutions.2 These adenocarcinomas have different biological and pathological features, as well as clinical course and overall prognosis.3 Therefore, failure to identify the correct site of origin may lead to considerable bias of long-term survival estimates.1–6 Survival estimates may also be biased if lymph node evaluation is inaccurate.7–10 Failure to identify regional

234

lymph node metastasis in true positive cases may result from insufficient surgical dissection or from inadequate histopathologic evaluation of the resected specimen. Evaluation of at least 15 lymph nodes has been suggested as a quality measure in staging of pancreatic cancer.8,9 According to current guidelines, adequate staging of nodenegative pancreatic cancer requires evaluation of minimum 12 lymph nodes, although N0 is assigned even when this number is not met as long as all evaluated lymph nodes are found negative.11 The ratio of the number of metastatic lymph nodes to the total number of lymph nodes evaluated, i.e., the lymph node ratio (LNR), may to some degree compensate for variations in surgical and pathological node dissection.12 LNR has thus been proposed as an alternative measure of nodal involvement in adenocarcinomas of the pancreatic head (PC), ampulla (AC), and distal bile duct (DBC).13–15 In cases for which positive lymph nodes have been detected, a high LNR would result from detection of either many positive nodes or identification of relatively few total nodes. However, provided sufficient lymph node dissection to determine adequately the nodal status, the actual number of positive lymph nodes could be an even better measure of the extent of nodal involvement. The aim of the present study was to compare N status, LNR, and the actual number of positive lymph nodes after pancreatoduodenectomy for AC, biliary, and pancreatic adenocarcinomas. Higher prevalence of lymph node involvement in pancreatic cancer compared to the two other origins, as well as other differences between periampullary cancers, may justify that the best method of lymph node staging could depend on the site of cancer origin.3 However, varying methods of tumor origin determination makes it difficult to draw conclusions from previous studies that have evaluated the different methods of nodal staging reporting from the separate anatomic locations. To our knowledge, this is the first study to compare N status, LNR, and count of metastatic lymph nodes among all three included origins in a setting where the tumor origin has been determined by standardized histopathological evaluation. PATIENTS AND METHODS The study included 179 patients (82 women and 97 men) who underwent a standard Whipple procedure with curative intent 1998–2009 at the Department of Surgery, Rikshospitalet. No patients underwent extended lymphadenectomy. Permission for the study was obtained from the Regional Committee for Research Ethics. Hospital records were reviewed retrospectively, and standard demographic, clinicopathological, and tumor-specific data were collected. Overall survival data were obtained from the National Population Registry in Norway, updated

E. Pomianowska et al.

February 8, 2012. All Norwegian inhabitants receive a unique personal identification number, enabling near complete follow-up. No patients were lost to follow-up in the present study. Patients were followed until death or censored after maximum 5 years (60 months). Thus, 144 patients were dead by end of the study, 34 patients were followed for the full length of 5 years, and median follow-up for the remaining 35 survivors was 87 months (interquartile range 55–129 months). Perioperative death (defined as death within 30 days of operation) was included in the association analyses, but excluded from survival analysis (AC, two patients; DBC, two patients; PC, no patients). Analysis including perioperative death gave very similar results. None of the patients received preoperative chemo- or radiotherapy. Until December 2007, national guidelines did not recommend adjuvant treatment. From 2008, adjuvant treatment with 5-fluorouracil was recommended for eligible patients resected from pancreatic cancer irrespective of nodal status, whereas radiotherapy was not recommended. For the two other origins, there were no national treatment guidelines regarding adjuvant treatment after pancreatoduodenectomy. HISTOPATHOLOGICAL EVALUATION OF RESECTION SPECIMENS The resection specimens were examined and handled according to a standardized protocol as described previously.16 The tumor origin was not always easy to assess, especially if the tumors were large with relation to several structures of possible origin. In such cases, dysplastic changes in ductal structures could be of importance for final classification. All tumors were allocated to one of the following locations: duodenum, ampulla of Vater, distal bile duct, or pancreas. Duodenal tumors were not included in this study.1 Nodal status was registered prospectively by routine pathologists according to the standardized protocol. Numbers of positive and total lymph nodes were extracted from the pathology reports, and LNR was calculated according to these numbers. The present study did not discriminate between nodal involvement because of direct invasion or true metastasis, in accordance with recommended principles for staging of nodal involvement.17 The following histopathologic factors were also prospectively registered: maximum tumor diameter, degree of differentiation, perineural infiltration, vascular infiltration, ductal dysplasia, lymph node status, and resection margins. According to the standardized protocol for histopathologic evaluation, the following resection margins were prospectively evaluated and retrospectively reevaluated: the pancreatic neck margin, the distal bile duct margin, and the retroperitoneal margin.16 All registrations of these parameters were reevaluated by an experienced pancreatic pathologist (OPFC).

LNR in Periampullary Adenocarcinomas

STATISTICAL ANALYSIS Associations between variables were examined by the Chi-square test, Fisher’s exact test, Mann–Whitney test, and Kruskal–Wallis test. Continuous variables were reported as median with corresponding range or interquartile range. Unadjusted survival analysis was performed by the Kaplan–Meier method, comparing curves by the log rank test. Multivariable Cox regression analysis was used for adjusted survival analysis. The proportional hazard assumption was tested graphically. Possible interactions were evaluated by inclusion of an interaction term in the models. For all tests, a two-sided p value of \0.05 was considered statistically significant. Statistical analyses were performed in SPSS 19 for Windows (SPSS, Chicago, IL). RESULTS During the study, 179 patients underwent pancreatoduodenectomy for adenocarcinomas originating from the ampulla (n = 61, 34 %), distal bile duct (n = 46, 26 %), or pancreatic head (n = 72, 40 %). Median age at time of resection was similar for the three groups (67 years, range 38–81; p = 0.59). Overall 5 year survival was 6 % for PC, 26 % for DBC, and 46 % for AC (p \ 0.001; Supplementary File 1). Histopathological characteristics are shown in Table 1 (details for R status and T classification in Supplementary File 2). PC were significantly larger (p \ 0.001) than both AC and DBC. Compared to PC, AC significantly more often had high degree of differentiation (p = 0.015), were more often resected with free margins (p \ 0.001), and had less often vascular (p = 0.002) or perineural (p \ 0.001) TABLE 1 Clinicopathological variables in 179 consecutive pancreatoduodenectomies for AC, DBC, and PC

235

invasion (Table 1). DBC had histopathological characteristics similar to PC, but were smaller (p \ 0.001) and had less frequent lymph node involvement. Lymph node involvement was more frequent in PC (n = 54, 75 %) compared to both AC (n = 29, 48 %; p = 0.001) and DBC (n = 26, 57 %, p = 0.036). The total number of lymph nodes evaluated was similar for all three origins in N0 resections (p = 0.91) as well as in N1 resections (p = 0.124) (Supplementary File 3). In unadjusted overall survival analysis, nodal involvement was significantly associated with poor survival for AC (p \ 0.001) and DBC (p \ 0.001) (Fig. 1a). However, an increasing number of metastatic lymph nodes did not seem to influence negatively on long-term overall survival (AC, p = 0.51; DBC, p = 0.09; Fig. 1b). Furthermore, higher LNR was not associated with poorer overall survival neither in AC (p = 0.17) nor in DBC (p = 0.19) (Fig. 1c). In PC (Fig. 2), there was no difference in overall survival comparing N0 and N1 resections (p = 0.31; Fig. 2a). In particular, there was no difference in overall survival for patients with 0 or 1 affected lymph node (median 17.5 and 25.4 months, respectively; p = 0.23). Furthermore, overall survival for patients with 2, 3, and C4 metastatic lymph nodes did not indicate that increasing numbers of involved lymph nodes affected overall prognosis (median survival 12.4, 17.7, and 13.4 months, respectively; p = 0.22; Fig. 2b). However, using a predefined cutoff threshold, LNR = 0.2, LNR significantly discriminated between patients with favorable versus nonfavorable overall survival among N0 and N1 resections (LNR B 0.2, median survival 17.5 months; LNR [ 0.2, median survival 12.4 months; p = 0.002) and among N1 resections alone (LNR B 0.2, median survival 15.2 months; LNR [ 0.2,

Characteristic

AC n = 61 (34 %)

DBC n = 46 (26 %)

PC n = 72 (40 %)

Tumor diameter (mm), median (IQR)

20 (16–27)

24 (16–32)

30 (25–40)

R1 resection status, n (%) Vascular invasion, n (%)

15 (25) 22 (36)

25 (54) 26 (57)

38 (53) 46 (64)

Perineural infiltration, n (%)

20 (33)

36 (78)

59 (82)

Grade I, II (%)

45 (74)

26 (56)

39 (54)

Grade III, IV (%)

16 (26)

20 (44)

10 (1–23)

13 (1–40)

11 (0–23)

1 (0–10)

2 (0–12)

Degree of differentiation

No. of examined nodes, median (range) No. of involved nodes, median (range) Lymph node metastasis (N1), n (%) AC ampullary adenocarcinoma, DBC distal bile duct adenocarcinoma, PC pancreatic adenocarcinoma, IQR interquartile range

0 (0–15) 29 (48)

26 (57)

33 (46)

54 (75)

Lymph node ratio, n (%) 0 (%)

32 (52)

20 (43)

17 (24)

[0–0.2 (%)

14 (23)

9 (20)

26 (36)

[0.2 (%)

15 (25)

17 (37)

29 (40)

236

FIG. 1 Unadjusted overall survival analysis for AC and DBC stratified by a lymph node status (p \ 0.001 both for AC and DBC), b number of lymph node metastases (p = 0.51 for AC and

E. Pomianowska et al.

p = 0.09 for DBC), and c LNR (LNR B 0.2 vs. LNR [ 0.2) (p = 0.17 for AC and p = 0.19 for DBC)

FIG. 2 Unadjusted overall survival analysis for PC stratified by a lymph node status (p = 0.32), b number of lymph node metastases (p = 0.22), and c LNR in node-positive resections (p = 0.011, comparing LNR B 0.2 vs. LNR [ 0.2)

median survival 12.4 months; p = 0.011; Fig. 2c). Finally, to examine the possible influence of understaging of nodal involvement, we compared N status and LNR when \12 vs. C12 lymph nodes had been removed. We found that LNR, but not N status, significantly discriminated between prognostic groups irrespective of the total number of lymph nodes removed. In particular, when \12 nodes

had been removed, LNR (Fig. 3b, p = 0.036), but not N status (Fig. 3a, p = 0.63), significantly predicted long-term patient survival. LNR may thus in particular better predict long-term outcome in patients with PC undergoing inadequate lymph node assessment. Univariate analysis thus indicated that N status evaluation was sufficient, and superior to both LNR and number of

LNR in Periampullary Adenocarcinomas

237

FIG. 3 Unadjusted overall survival analysis for nodal status and LNR in PC, stratified by total number of lymph nodes removed,\12 vs. C12. a N1 vs. N0, lymph node yield \12 (p = 0.63). b LNR [ 0.20 vs. LNR B 0.20, lymph node yield \12 (p = 0.036). c N1 vs. N0, lymph node yield C12 (p = 0.27). d LNR [ 0.20 vs. LNR B 0.20, lymph node yield C12 (p = 0.014)

metastatic nodes, in nodal evaluation for AC and DBC, whereas LNR might be superior for evaluation of PC. Consequently, we used N status to evaluate the independent role of nodal involvement in adjusted survival analysis for AC (hazard ratio [HR] 3.54; confidence interval [CI] 1.43–8.76; p = 0.01) and DBC (HR 3.38 (1.42–8.05; p = 0.01) (Fig. 4a). Next, we compared LNR and N status in two separate models in the adjusted analysis for PC (Fig. 4b), in order to confirm whether LNR may in fact be the more powerful prognostic indicator among these patients. We found that LNR [0.2 independently predicted a poor prognosis adjusting for R status, degree of differentiation, and tumor diameter (p = 0.032; HR 1.87, 95 % CI 1.06–3.31; model 1), whereas N status was not significantly associated with a poor prognosis in PC (p = 0.52; HR 1.23, 95 % CI 0.65–2.33; model 2).

long-term survival after pancreatoduodenectomy for PC, DBC, and AC. Reported frequencies for each of these sites vary considerably.1 Nonstandardized evaluation of the cancer origin may lead to misclassification of the tumor origin and in turn bias prognostic estimates.1–6 Previous studies comparing the different ways of characterizing lymph node involvement have either reported on each of these locations separately (PC, DBC, and AC), or without slide review to ascertain the cancer origin.13–15,18–36 There is conflicting evidence as to which method is the better for evaluation of nodal metastasis in adenocarcinomas arising from the separate anatomic locations (Table 2). The present study demonstrates that N status per se is sufficient to evaluate adequately nodal involvement in AC and DBC, whereas LNR may be the more powerful indicator of long-term survival in PC.

DISCUSSION

Nodal Evaluation in PC

The present study aimed to ascertain whether N status, LNR, or number of metastatic lymph nodes better predicts

Berger et al. were the first to compare LNR, N status, and count of metastatic nodes as prognostic indicators after

238 FIG. 4 Adjusted overall survival analysis for a AC and DBC; and for b PC, comparing N status in model 1 and LNR in model 2

E. Pomianowska et al.

a

AC (N=61)

R status (R1 vs R0) Grade (low vs high) Diameter (split 2 cm) N status (N1 vs N0)

DBC (N=46) R status (R1 vs R0) Grade (low vs high) Diameter (split 2 cm) N status (N1 vs N0)

b Model 1 R status (R1 vs R0) Grade (low vs high) Diameter (split 2 cm)

LNR (>0.2 vs 0.2)

PC (N=72)

p-value 0 .0 2 0.17 0.42 0.002

HR 2.87 1.81 1.48 4.98

95% CI 1.19 - 6.94 0.78 - 4.18 0.57 - 3.82 1.83 - 13.6

p-value 0 .0 3 0.49 0.59 0 .0 1

HR 2.49 1.30 1.23 3.15

95% CI 1.11 5.58 0.62 2.76 0.58 2.57 1.32 7.53

p-value 0.63 0.12 0.29 0 .0 3

HR 1.14 1.51 1.52 1.87

95% CI 0.68 1.89 0.90 2.51 0.70 3.27 1.06 3.31

0.55 0 .0 3 0.12 0.52

1.18 1.75 1.87 1.23

0.69 1.06 0.86 0.65

Model 2 R status (R1 vs R0) Grade (low vs high) Diameter (split 2 cm) N status (N1 vs N0)

resection for PC.13 They found no difference comparing LNR = 0 with LNR B 0.15, whereas LNR [ 0.15 significantly predicted poor long-term survival. However, this study did not separately compare survival between LNR [ 0.15 and LNR B 0.15 among node-positive resections. Sierzega et al. 18 confirmed that patients with high LNR (using 0.20 as their cut-point) had unfavorable long-term survival compared to patients with low LNR among node-positive resections. In the present study, we defined a LNR value of 0.20 as the cutpoint a priori, thus avoiding overestimation of prognostic values.30,37 In most previous studies (Table 2), the cut-point for LNR used to define subsets of favorable versus nonfavorable survival has been between 0.15 and 0.20 (Supplementary File 4). A study designed specifically to determine the best cut-point for LNR concluded that this might be 0.18 in PC.20 However, the mean number of evaluated lymph nodes was 17 in that study, and the definition of LNR (ratio of metastatic to total nodes) implies that the optimal cut-point depends on the average number of evaluated lymph nodes. A lower average number of evaluated lymph nodes in some studies may indicate understaging.20,27, 30,38 In the present study, lymph nodes were obtained by en bloc dissection of primary metastatic

-

2.02 2.88 4.08 2.33

regions, not including secondary sites such as the periaortic lymph nodes, and with standardized evaluation of the resected specimen.39–42 We acknowledge that our study reports a lower median number of lymph nodes (eleven nodes) compared to the recommended minimum number (twelve nodes), and that this may be a limitation to the present study.11 However, the median number of evaluated lymph nodes is most frequently B9 even in academic institutions, and a high lymph node yield could also in some reports result from resections involving extended lymphadenectomy.9,43 The median number of lymph nodes evaluated in the present study is thus representative of a medium-volume tertiary referral institution. N status is insufficient as prognostic indicator in PC specifically because involvement of 0 or 1 lymph node does not make a difference to overall survival (Fig. 2b). This has also been reported previously in three separate studies, regardless of whether the median number of total lymph nodes examined was higher or lower than 11.23,25,30 Increasing lymph node involvement, beyond one affected node, has been shown by many to be inversely associated with long-term survival in PC.19–21,23,25,27,30 Involvement of one lymph node may in some cases be a consequence of direct lymph node invasion rather than true metastasis.

LNR in Periampullary Adenocarcinomas

239

TABLE 2 LNR and N status in pancreatoduodenectomy reports on pancreatic, distal bile duct and ampullary adenocarcinomas Study

n

N1 (%)

Median lymph node count

Prognostic factor

n

N status

Range

Total LNR

N1 LNR

Pancreatic adenocarcinomas Berger13

129a

64

17

2–55

NR

Yes

NR

Sierzega18

96

67

17

5–92

Yes

Yes

Yes

Pawlik19

905

79

17

6–28

Yes

Yes

Yes

House20

696b

65

17c

NR

Yes

Yes

NR

Slidel21

3868a,b

51

7

0–90

Yes

Yes

Yes

Smith22 Riediger23

109 182b

81 70

17 16

NR 2–47

Yes No

Yes Yes

NR Yes

Hurtuk15

219

73

18

1–45

Yes

Yes

NR

Bhatti24

84

67

9

1–26

Yes

Yes

Yes

Murakami25

119b

66

28

2–75

Yes

Yes

No

La Torre26

101b

63

20c

3–26

Yes

Yes

NR

Showalter27

445b,d

53

9–11

1–56

Yes

Yes

Yes

Robinson28

134

90

19

7–36

Yes

Yes

NR

Sanjay29

51

78

20

0–41

No

Yes

NR

Huebner30

499a

53

10

1–74

Yes

Yes

Yes

This study

72

75

11

0–23

No

Yes

Yes

Distal bile duct adenocarcinomas Hurtuk15

57

14

2–36

Yes

No

NR

Kawai

b

62

39

12

5–38

Yes

Yes

NR

Oshiro33

60b

43

14

0–57

Yes

Yes

NR

57

13

1–40

Yes

Yes

No

31

35

This study 46 Ampullary adenocarcinomas Falconi14

90

50

16

5–47

Yes

Yes

NR

15

Hurtuk

75

41

15

2–38

Yes

Yes

NR

Sierzega34

111b

47

15

4–62

Yes

Yes

NR

Lee35

52

39

26

10–60

Yes

Yes

NR

Sakata36

71

48

28

NR

Yes

Yes

NR

This study

61

48

10

1–23

Yes

Yes

No

N status nodal status, LNR lymph node ratio, LNR (tot) LNR in univariate overall survival analysis including all patients (N0 and N1 resections), LNR (N1) LNR in univariate overall survival analysis including only N1 resections, NR not reported a Patients with incomplete reports on lymph nodes were excluded b

Including not only pancreatoduodenectomy but also, for example, distal pancreatectomy and hepatectomy

c

Mean

d

Patients for whom no lymph nodes had been removed were excluded

Importantly, overall survival for patients with direct invasion into regional lymph nodes is similar compared to overall survival for patients with node-free resections.44 In current staging recommendations, and in the present study, N1 is assigned irrespective of whether nodal involvement was the result of direct invasion or true metastasis.17 LNR may adjust to some degree for the vast ranges of total numbers of lymph nodes evaluated in pancreatoduodenectomy reports (Table 2). Pawlik et al. 19 demonstrated that LNR is better able to discriminate prognosis in subgroups of patients with similar numbers of positive lymph nodes. The more aggressive biology of PC compared to AC

and DBC may also explain why involvement of one versus no lymph nodes is associated with similar long-term survival in PC.3 Nodal Evaluation in Distal Bile Duct and AC Hurtuk et al. 15 were the first to evaluate LNR in DBC after pancreatoduodenectomy. Two further studies evaluated LNR in extrahepatic bile duct cancer although not limited to pancreatoduodenectomies (Table 2).31,33 These studies indicate that increasing LNR may be associated with worse overall survival, although high LNR was

240

compared to low LNR including node-negative resections in the latter category. This does not prove that LNR adds prognostic information to that which is provided by N status evaluation. It could be mentioned that in PC, the reference to LNR = 0 when comparing increasing LNR is appropriate because the prognostic impact of 0 or 1 lymph node metastasis does not differ. LNR has not been evaluated previously among nodepositive pancreatoduodenectomies for DBC or AC. The present study was unable to demonstrate that increasing LNR is significantly associated with worse overall survival among patients with node-positive DBC or AC. Whereas most studies in AC and DBC reported median survival for node-negative patients more than 5 years, only two studies (both in AC) reported median survival for node-positive patients more than 2 years (Supplementary File 4).35,36 Traditional nodal evaluation (N0 vs. N1) is thus sufficient to stratify patients in distinct prognostic subgroups after pancreatoduodenectomy for AC and DBC. However, because of our relatively small sample size, larger studies must be performed to clarify whether increasing LNR could in fact have some importance for long-term survival among these patients. Conclusion The predictive value of nodal involvement depends on the cancer origin. In AC and DBC, nodal status distinctly predicts long-term survival. In PC, LNR may discriminate better between prognostic groups than nodal status or count of metastatic nodes. In particular, LNR is more predictive of long-term outcome if lymph node assessment is insufficient. These findings may have implications for patient selection and stratification in future clinical trials on adjuvant therapies in the separate cancer entities. REFERENCES 1. Verbeke CS, Gladhaug IP. Resection margin involvement and tumour origin in pancreatic head cancer. Br J Surg. In press. 2. Westgaard A, Laronningen S, Mellem C, et al. Are survival predictions reliable? Hospital volume versus standardisation of histopathologic reporting for accuracy of survival estimates after pancreatoduodenectomy for adenocarcinoma. Eur J Cancer. 2009;45:2850–9. 3. Sarmiento JM, Nagomey DM, Sarr MG, Farnell MB. Periampullary cancers: are there differences? Surg Clin North Am. 2001;81:543–55. 4. Carpelan-Holmstrom M, Nordling S, Pukkala E, et al. Does anyone survive pancreatic ductal adenocarcinoma? A nationwide study re-evaluating the data of the finnish cancer registry. Gut. 2005;54:385–7. 5. Westgaard A, Tafjord S, Farstad IN, et al. Pancreatobiliary versus intestinal histologic type of differentiation is an independent prognostic factor in resected periampullary adenocarcinoma. BMC Cancer. 2008;8:170.

E. Pomianowska et al. 6. Westgaard A, Clausen OPF, Gladhaug IP. Survival estimates after pancreatoduodenectomy skewed by non-standardized histopathology reports. APMIS. 2011;119:689–700. 7. Feinstein AR, Sosin DM, Wells CK. The Will Rogers phenomenon. Stage migration and new diagnostic techniques as a source of misleading statistics for survival in cancer. N Engl J Med. 1985;312:1604–8. 8. Tomlinson JS, Jain S, Bentrem DJ, et al. Accuracy of staging node-negative pancreas cancer: a potential quality measure. Arch Surg. 2007;142:767–823. 9. Bilimoria KY, Talamonti MS, Wayne JD, et al. Effect of hospital type and volume on lymph node evaluation for gastric and pancreatic cancer. Arch Surg. 2008;143:671–8. 10. Adsay NV, Basturk O, Altinel D, et al. The number of lymph nodes identified in a simple pancreatoduodenectomy specimen: comparison of conventional vs orange-peeling approach in pathologic assessment. Mod Pathol. 2009;22:107–12. 11. Exocrine and endocrine pancreas. In: Edge SB, Byrd DR, Compton CC, Fritz AG, Green FL, Trotti A (eds). AJCC cancer staging manual. 7 edition. Chicago: Springer; 2010. p. 241–9. 12. Petrelli F, Borgonovo K, Barni S. The emerging issue of ratio of metastatic to resected lymph nodes in gastrointestinal cancers: an overview of literature. Eur J Surg Oncol. 2011;37:836–47. 13. Berger AC, Watson JC, Ross EA, Hoffman JP. The metastatic/ examined lymph node ratio is an important prognostic factor after pancreaticoduodenectomy for pancreatic adenocarcinoma. Am Surg. 2004;70:235–40. 14. Falconi M, Crippa S, Dominguez I, et al. Prognostic relevance of lymph node ratio and number of resected nodes after curative resection of ampulla of Vater carcinoma. Ann Surg Oncol. 2008;15:3178–86. 15. Hurtuk MG, Hughes C, Shoup M, Aranha GV. Does lymph node ratio impact survival in resected periampullary malignancies? Am J Surg. 2009;197:348–52. 16. Westgaard A, Tafjord S, Farstad IN, et al. Resectable adenocarcinomas in the pancreatic head: the retroperitoneal resection margin is an independent prognostic factor. BMC Cancer. 2008;8:5. 17. Purposes and principles of cancer staging. In: Edge SB, Byrd DR, Compton CC, Fritz AG, Green FL, Trotti A (ed). AJCC cancer staging manual. 7 ed. Chicago: Springer; 2010. p. 3–14. 18. Sierzega M, Popiela T, Kulig J, Nowak K. The ratio of metastatic/resected lymph nodes is an independent prognostic factor in patients with node-positive pancreatic head cancer. Pancreas. 2006;33:240–5. 19. Pawlik TM, Gleisner AL, Cameron JL, et al. Prognostic relevance of lymph node ratio following pancreaticoduodenectomy for pancreatic cancer. Surgery. 2007;141:610–8. 20. House MG, Gonen M, Jarnagin WR, et al. Prognostic significance of pathologic nodal status in patients with resected pancreatic cancer. J Gastrointest Surg. 2007;11:1549–55. 21. Slidell MB, Chang DC, Cameron JL, et al. Impact of total lymph node count and lymph node ratio on staging and survival after pancreatectomy for pancreatic adenocarcinoma: a large, population-based analysis. Ann Surg Oncol. 2008;15:165–74. 22. Smith RA, Bosonnet L, Ghaneh P, et al. Preoperative CA19-9 levels and lymph node ratio are independent predictors of survival in patients with resected pancreatic ductal adenocarcinoma. Dig Surg. 2008;25:226–32. 23. Riediger H, Keck T, Wellner U, et al. The lymph node ratio is the strongest prognostic factor after resection of pancreatic cancer. J Gastrointest Surg. 2009;13:1337–44. 24. Bhatti I, Peacock O, Awan AK, Semeraro D, Larvin M, Hall RI. Lymph node ratio versus number of affected lymph nodes as predictors of survival for resected pancreatic adenocarcinoma. World J Surg. 2010;34:768–75.

LNR in Periampullary Adenocarcinomas 25. Murakami Y, Uemura K, Sudo T, et al. Number of metastatic lymph nodes, but not lymph node ratio, is an independent prognostic factor after resection of pancreatic carcinoma. J Am Coll Surg. 2010;211:196–204. 26. La Torre M, Cavallini M, Ramacciato G, et al. Role of the lymph node ratio in pancreatic ductal adenocarcinoma. Impact on patient stratification and prognosis. J Surg Oncol. 2011;104: 629–33. 27. Showalter TN, Winter KA, Berger AC, et al. The influence of total nodes examined, number of positive nodes, and lymph node ratio on survival after surgical resection and adjuvant chemoradiation for pancreatic cancer: a secondary analysis of RTOG 9704. Int J Radiat Oncol Biol Phys. 2011;81:1328–35. 28. Robinson SM, Rahman A, Haugk B, et al. Metastatic lymph node ratio as an important prognostic factor in pancreatic ductal adenocarcinoma. Eur J Surg Oncol. 2012;38:333–9. 29. Sanjay P, de Figueiredo RS, Leaver H, et al. Preoperative serum C-reactive protein levels and post-operative lymph node ratio are important predictors of survival after pancreaticoduodenectomy for pancreatic ductal adenocarcinoma. JOP. 2012;13:199–204. 30. Huebner M, Kendrick M, Reid-Lombardo KM, et al. Number of lymph nodes evaluated: prognostic value in pancreatic adenocarcinoma. J Gastrointest Surg. In press. 31. Kawai M, Tani M, Kobayashi Y, et al. The ratio between metastatic and examined lymph nodes is an independent prognostic factor for patients with resectable middle and distal bile duct carcinoma. Am J Surg. 2010;199:447–52. 32. Murakami Y, Uemura K, Hayashidani Y, Sudo T, Ohge H, Sueda T. Pancreatoduodenectomy for distal cholangiocarcinoma: prognostic impact of lymph node metastasis. World J Surg. 2007;31:337–42. 33. Oshiro Y, Sasaki R, Kobayashi A, et al. Prognostic relevance of the lymph node ratio in surgical patients with extrahepatic cholangiocarcinoma. Eur J Surg Oncol. 2011;37:60–4. 34. Sierzega M, Nowak K, Kulig J, Matyja A, Nowak W, Popiela T. Lymph node involvement in ampullary cancer: the importance of

241

35.

36.

37.

38.

39.

40.

41. 42.

43.

44.

the number, ratio, and location of metastatic nodes. J Surg Oncol. 2009;100:19–24. Lee JH, Lee KG, Ha TK, et al. Pattern analysis of lymph node metastasis and the prognostic importance of number of metastatic nodes in ampullary adenocarcinoma. Am Surg. 2011;77:322–9. Sakata J, Shirai Y, Wakai T, Ajioka Y, Akazawa K, Hatakeyama K. Assessment of the nodal status in ampullary carcinoma: the number of positive lymph nodes versus the lymph node ratio. World J Surg. 2011;35:2118–24. Altman DG, Lausen B, Sauerbrei W, Schumacher M. Dangers of using ‘‘optimal’’ cutpoints in the evaluation of prognostic factors. J Natl Cancer Inst. 1994;86:829–35. Schwarz RE, Smith DD. Extent of lymph node retrieval and pancreatic cancer survival: information from a large U.S. population database. Ann Surg Oncol. 2006;13:1189–200. Michalski CW, Kleeff J, Wente MN, Diener MK, Buchler MW, Friess H. Systematic review and meta-analysis of standard and extended lymphadenectomy in pancreaticoduodenectomy for pancreatic cancer. Br J Surg. 2007;94:265–73. Iqbal N, Lovegrove RE, Tilney HS, et al. A comparison of pancreaticoduodenectomy with extended pancreaticoduodenectomy: a meta-analysis of 1909 patients. Eur J Surg Oncol. 2009;35:79–86. Reddy SK, Tyler DS, Pappas TN, Clary BM. Extended resection for pancreatic adenocarcinoma. Oncologist. 2007;12:654–63. Kennedy EP, Yeo CJ. Pancreaticoduodenectomy with extended retroperitoneal lymphadenectomy for periampullary adenocarcinoma. Surg Oncol Clin N Am. 2007;16:157–76. Rupp CC, Linehan DC. Extended lymphadenectomy in the surgery of pancreatic adenocarcinoma and its relation to quality improvement issues. J Surg Oncol. 2009;99:207–14. Pai RK, Beck AH, Mitchem J, et al. Pattern of lymph node involvement and prognosis in pancreatic adenocarcinoma: direct lymph node invasion has similar survival to node-negative disease. Am J Surg Pathol. 2011;35:228–34.