ORIGINAL E n d o c r i n e
ARTICLE C a r e
Prognostic Value of Microscopic Lymph Node Involvement in Patients With Papillary Thyroid Cancer Stéphane Bardet, Renaud Ciappuccini, Elske Quak, Jean-Pierre Rame, David Blanchard, Dominique de Raucourt, Emmanuel Babin, Jean-Jacques Michels, Dominique Vaur, and Natacha Heutte Department of Nuclear Medicine and Thyroid Unit (S.B., R.C., E.Q.), and Departments of Head and Neck Surgery (J.-P.R., D.B., D.d.R.), and Pathology (J.-J.M.), Biology (D.V.), Centre François Baclesse, 14076 Caen, France; Department of Head and Neck Surgery (E.B.), University Hospital, Caen 14000, France; Unité 1086 (N.H.), INSERM-University of Caen-Basse Normandie, “Cancers and Préventions” Program, University of Caen-Basse Normandie, 14032 Caen, France
Context: The impact of microscopic nodal involvement on the risk of persistent/recurrent disease (PRD) remains controversial in patients with papillary thyroid carcinoma (PTC). Objective: The goal of the study was to assess the risk of PRD and the 4-year outcome in PTC patients according to their initial nodal status [pNx, pN0, pN1 microscopic (cN0/pN1) or pN1 macroscopic (cN1/pN1)]. Design: We conducted a retrospective cohort study. Patients: The study included 305 consecutive PTC patients referred for radioiodine ablation from 2006 to 2011. Main Outcome Measure: We evaluated the risk of structural PRD and the disease status at the last follow-up. At ablation, persistent disease was consistently assessed by using post-radioiodine ablation scintigraphy combining total body scan and neck and thorax single-photon computed tomography-computed tomography (SPECT-CT) acquisition. Results: Of 305 patients, 128 (42%) were pNx, 84 (28%) pN0, 44 (14%) pN1 microscopic, and 49 (16%) pN1 macroscopic. The 4-year cumulative risk of PRD was higher in pN1 macroscopic than in pN1 microscopic patients (49% vs 24%, P ⫽ .03), and higher in pN1 microscopic than in pN0 (12%, P ⫽ .01) or pNx patients (6%, P ⬍ .001). On multivariate analysis, tumor size of 20 mm or greater [relative risk (RR) 3.4; P ⫽ .0001], extrathyroid extension (RR 2.6; P ⬍ .003), pN1 macroscopic (RR 4.5; P ⬍ .0001), and pN1 microscopic (RR 2.5; P ⬍ .02) were independent risk factors for PRD. At the last visit, the proportion of patients with no evidence of disease decreased from pNx (98%), pN0 (93%), and pN1 microscopic (89%) to pN1 macroscopic patients (70%) (P ⬍ .0001, Cochran-Armitage trend test). Extrathyroid extension (odds ratio 9.7; P ⬍ .0001) and N1 macroscopic (OR 4.9; P ⬍ .001) independently predicted persistent disease at the last visit, but N1 microscopic did not. Conclusions: Patients with microscopic lymph node involvement present an intermediate outcome between that observed in pN0-pNx patients and pN1 macroscopic patients. These data may justify modifications to the risk recurrence staging systems. (J Clin Endocrinol Metab 100: 132–140, 2015)
ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2015 by the Endocrine Society Received January 16, 2014. Accepted September 30, 2014. First Published Online October 10, 2014
132
jcem.endojournals.org
Abbreviations: AJCC, American Joint Committee on Cancer; ECE, extracapsular extension; FDG-PET-CT, 18F-fluorodeoxyglucose-positron emission tomography-computed tomography; LND, lymph node dissection; NED, no evidence of disease; OR, odds ratio; PRD, persistent/recurrent disease; PTC, papillary thyroid carcinoma; RAI, radioiodine; rhTSH, recombinant human TSH RR, relative risk; SPECT-CT, single-photon emission computed tomography-computed tomography; Tg, thyroglobulin; TgAb, Tg antibody; TNM, tumor node metastasis; US, ultrasound.
J Clin Endocrinol Metab, January 2015, 100(1):132–140
doi: 10.1210/jc.2014-1199
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
doi: 10.1210/jc.2014-1199
he presence of nodal metastases on the primary pathological report (pN1) is known to increase the risk of locoregional persistent/recurrent disease (PRD) and, in some instances, the risk of disease-specific mortality in patients with papillary thyroid carcinoma (PTC). The magnitude of risk depends, however, on several characteristics of the nodal involvement, mainly its volume (1). Schematically, large-volume, macroscopic, clinically apparent nodal involvement has been associated with rates of recurrence of 20%–25% (2– 4). Conversely, small-volume, microscopic nodal involvement, has been related to rates below 5% (2, 4), not clearly higher than those estimated in patients without nodal involvement (pN0) after prophylactic lymph node dissection (LND) or unknown nodal status (pNx). However, the presence of nodal involvement, regardless of size, has an upstaging impact in both the American Joint Committee on Cancer (AJCC) tumor node metastasis (TNM) staging system (5) and the American Thyroid Association risk of recurrence system (6). In a previous study on 545 PTC patients diagnosed between 1965 and 2003, of whom 91% have had radioiodine (RAI), the 10-year cumulative risk of lymph node recurrence was estimated at 7% with rates at 4% in pNx, 2% in pN0, 4% in pN1 microscopic, and 19% in pN1 macroscopic patients (2). Macroscopic, but not microscopic, lymph node involvement was found as an independent risk factor of lymph node recurrence as well as extrathyroidal invasion and male gender. In the last 10 years, in addition to ultrasound (US), new functional imaging modalities have emerged improving the detection of small (⬍10 mm) lesions, in or outside the neck, which can be either not visible or considered as benign on conventional radiological modalities. These include single-photon emission computed tomography-computed tomography (SPECT-CT) imaging during 131I scintigraphy, mainly in the postablation setting (7–9), and 18F-fluorodeoxyglucose-positron emission tomography-computed tomography (FDG-PET-CT) scan in the context of high serum thyroglobulin (Tg) levels, especially in iodine-refractory patients (10). The aim of the present study was therefore to reassess the risk of structural persistent/recurrent locoregional and distant disease in newly diagnosed PTC patients since 2006, treated with surgery and postoperative RAI, and to reexamine the impact of microscopic or macroscopic lymph node involvement on patient outcome.
T
jcem.endojournals.org
133
and subsequently followed up at our institute, were initially included in the present study. Three patients with a poorly differentiated component were excluded and 305 patients were retained for final analysis.
Surgery Surgery was performed either at our institution or at the university hospital (n ⫽ 199) or in other institutions (n ⫽ 106). Most patients (n ⫽ 264) were operated on by head and neck surgeons. Preoperative neck US for the assessment of thyroid nodules was performed by radiologists working inside or outside our institution. All patients underwent a total thyroidectomy in one (n ⫽ 233) or two interventions (n ⫽ 72). The decision to perform or not to perform a prophylactic or a therapeutic LND, and its extent, depended on several factors including the cN0/cN1 distinction of the AJCC TNM staging system (5), the surgeon’s preference or experience, and the preoperative cytological and the frozen section examination data when performed. In cN0 patients, surgery included either a prophylactic LND of the central compartment (level VI) sometimes associated with LND of the lateral compartments (levels IIa, III, and IV), or no LND at all. Central LND was bilateral in most cases, could be ipsilateral to the thyroid tumor only when frozen section examination was dubious, and rarely only contralateral when it was falsely negative, leading to total thyroidectomy in two interventions. In cN1 patients, a therapeutic LND was performed generally consisting of a central and lateral LND. The criteria applied for defining initial macroscopic or microscopic lymph node involvement have been previously reported (2). A lymph node metastasis was considered macroscopic (cN1/pN1 for the AJCC) when evident on clinical examination or US before surgery and/or clearly suspected by the surgeon according to his report. All other lymph node metastases were considered microscopic (cN0/pN1).
Pathology Diagnosis of PTC was based on the written pathological report. Pathological variants were defined according to the World Health Organization classification (11). Based on the pathological report, the number of resected and metastatic lymph nodes and the presence of nodal extracapsular extension (ECE) were notified for each patient. When available, the slides were reviewed to estimate the size of the largest metastatic nodal focus.
Radioiodine ablation RAI was administered either after thyroid hormone withdrawal or after recombinant human TSH (rhTSH) (0.9 mg, im, for 2 consecutive days; Thyrogen; Genzyme). The RAI activity (1.1 or 3.7 GBq) and the preparation modalities were determined after local multidisciplinary staff except for the 42 low-risk PTC patients randomized in the trial (12). A post-RAI total body scan was performed 2 or 5 days after 1.1 or 3.7 GBq, respectively, with SPECT-CT acquisition in all patients, at least on the neck and thorax, as previously described (7).
Tg and Tg antibody (TgAb) assay
Patients and Methods Patients Three hundred eight consecutive PTC patients referred to our department for RAI between January 2006 and December 2011,
Blood samples for stimulated serum Tg and TgAb measurements were collected immediately before RAI administration, and stored at ⫺20°C. All serum Tg levels were measured using Roche Cobas 6000 Tg (Roche Diagnostics) with a detection limit of 0.1 ng/mL and a functional sensitivity of 1.0 ng/mL. TgAb was
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
134
Bardet et al
Microscopic Lymph Node Involvement
measured using Immulite TgAb (Diagnostic Product Corp) from January 2006 to June 2008 and thereafter using Roche Cobas 6000 anti-Tg assay (Roche Diagnostics).
Definition of structural persistent/recurrent disease As previously proposed (13), structural PRD was defined as evidence of tumor in the thyroid bed, lymph node metastases or distant metastases after completion of initial treatment (ie, surgery and RAI) confirmed by either histology or complementary imaging modalities. At ablation, persistent disease was consistently assessed using the post-RAI scintigraphy combining total body scan and at least neck and thorax SPECT-CT. In the presence of scintigraphic abnormalities, other imaging modalities such as neck US, neck and thorax CT scan with contrast medium injection, FDG-PET-CT, or bone magnetic resonance imaging was performed at the discretion of the clinician to provide additional evidence of residual disease and to better define the disease extent. For patients with normal postablation scintigraphy but persistent detectable serum Tg level (⬎1 ng/mL), similar imaging modalities were performed to evidence RAI refractory lesions. At that point, we chose not to consider biochemical disease, ie, elevated serum Tg level (suppressed Tg ⱖ 1 ng/mL and/or stimulated Tg ⱖ 2 ng/mL) with no other evidence of disease. We arbitrarily defined persistent disease when the diagnosis was made within 1 year after the initial treatment and recurrent disease after this period. To assess tumor burden, PRD was classified into two categories. Small-volume disease was defined either by the presence of clearly abnormal foci on post-RAI scintigraphy or on FDGPET-CT without obvious lesions on conventional radiology or by small (⬍10 mm) but highly suspicious lesions on conventional radiology (neck US, CT scan, magnetic resonance imaging). Conversely, large-volume disease was defined by large (⬎10 mm) lesions clearly evidenced by conventional radiology, whatever the presence of scintigraphic abnormalities. The RAI refractory feature of the PRD was also specified as previously defined (14).
Clinical outcome assessment As previously described (7, 12), patients with a normal postablation scan were seen at 3 months on levothyroxine (LT4) treatment for serum TSH, Tg, and TgAb determinations. When suppressed Tg level was less than 1 ng/mL in the absence of TgAb, disease status was assessed at 9 months by serum rhTSH-stimulated Tg determination and neck US. In the presence of TgAb, a diagnostic 131I scintigraphy after rhTSH was also performed. When evaluation at 9 months was normal (ie, stimulated Tg ⱕ 1 ng/mL and normal US or post-131I scintigraphy), patients were annually followed up with suppressed serum Tg and TgAb measurements. For patients with abnormal postablation scintigraphy, the clinical management depended on the location and size of persistent disease. In patients with nodal disease, surgery was favored for large-volume lesions, whereas close surveillance or additional RAI treatment was preferred for small-volume disease. For distant metastases, RAI treatment was generally repeated at 6-month intervals until the disappearance of abnormal foci. Neck irradiation was considered in patients with inoperable local extension. For patients with RAI refractory lesions, close surveillance or therapeutic options were discussed.
J Clin Endocrinol Metab, January 2015, 100(1):132–140
At the last follow-up visit, the disease status was categorized into one of the following three subgroups (13): 1) no evidence of disease (NED) defined by a suppressed serum Tg level less than 1 ng/mL and, if performed, normal imaging; 2) persistent structural disease defined by a Tg level of 1 ng/mL or greater associated with imaging abnormalities; and 3) persistent biochemical disease with a Tg level of 1 ng/mL or greater and no other evidence of disease. In the presence of TgAb, disease status assessment was based on imaging data and serum Tg if there was a level of 1 ng/mL or greater.
Data analysis Patient characteristics were compared using the 2 test, the Fisher’s exact test, or the Kruskal-Wallis test, as appropriate. The trend test of Cochran-Armitage was used to examine proportions of disease-related events according to N status in the following subgroups: Nx, N0, N1 microscopic, and N1 macroscopic. Time to PRD was calculated from the day of postablation scintigraphy to the date of diagnosis of PRD or February first 2013, whichever came first. The date of diagnosis was either the date of biopsy or surgery or that of abnormal imaging modalities or that of postablation scintigraphy when the persistent tumor was demonstrated on scintigraphy only. In this latter case, time to persistent disease was arbitrarily defined as 1 month. The Kaplan-Meier method was used to estimate the probability of PRD. The estimates in each group were compared using the logrank test. The analysis of prognostic factors was performed using the multivariate Cox regression model for PRD and using logistic regression with stepwise selection for persistent structural or biochemical disease at the end of follow-up. Statistical significance was defined as P ⬍ .05. All tests were two sided. SAS 9.3 statistical software (SAS Institute) was used to analyze the data.
Results Patient characteristics The study group included 305 patients with PTC (238 women and 67 men, mean age 49 ⫾ 17 y). There were 185 patients with a conventional form of PTC (61%), 91 with a follicular variant (30%), five with a tall-cell or a columnar-cell variant (2%), seven with a diffuse sclerosing variant (2%), three with an oncocytic variant (1%), and 14 with a solid variant (4%). After surgery, 128 patients (42%) were considered pNx, 84 (28%) pN0, 44 (14%) pN1 microscopic, and 49 (16%) pN1 macroscopic (Table 1). The four N subgroups did not differ for age, sex ratio, aggressive pathological subtypes, bilateral tumors, RAI activity, and median follow-up but differed for tumor size (P ⬍ .001), multifocal tumors (P ⬍ .01), T status of the TNM classification (P ⬍ .0001), and extrathyroid extension (P ⬍ .0001). The LND modalities did not significantly differ between pN0 and pN1 microscopic patients (P ⫽ .09) with prophylactic central LND performed in 85% vs 73% and central and lateral LND in 14% vs 23% of patients, respectively. In con-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
doi: 10.1210/jc.2014-1199
Table 1.
jcem.endojournals.org
135
Characteristics of Patients According to Initial pN Status
Mean age ⫾SD, y Sex ratio (F/M) Aggressive pathological subtypesa Mean tumor size ⫾ SD, mm Multifocal tumor Bilateral tumor T status (TNM) T1a T1b T2 T3 T4 Extrathyroid extension Known distant metastases before surgery Dissection modalities No dissection Central dissection only Bilateral Ipsilateral Contralateral Lateral dissection only (ipsilateral) C and L dissection Bilateral C and L Bilateral C and Ipsilateral L Adenectomy N1a/N1b Mean number of resected LN ⫾ SD Mean number of metastatic LN ⫾ SD LN ECE Median size of the largest metastatic LN, mm Mean I131 activity ⫾ SD, MBq Preparation with THW Median follow-up, mo
pNx (n ⴝ 128)
pN0 (n ⴝ 84)
pN1 Microscopic (n ⴝ 44)
pN1 Macroscopic (n ⴝ 49)
P Value
52 ⫾ 14 3.6 (101/27) 7 (5%)
48 ⫾ 16 4.2 (68/16) 8 (10%)
47 ⫾ 18 4.0 (36/8) 5 (11%)
47 ⫾ 21 1.9 (33/16) 9 (18%)
.18 .25 .07
17 ⫾ 13
23 ⫾ 10
19 ⫾ 12
25 ⫾ 24
⬍.001
75 (59%) 51 (40%)
29 (35%) 25 (30%)
21 (48%) 16 (36%)
20 (41%) 16 (33%)
⬍.01 .48
48 (38%) 39 (31%) 30 (23%) 10 (8%) 1 (1%) 8 (6%) 1 (1%)
9 (11%) 28 (33%) 31 (37%) 14 (17%) 2 (2%) 13 (15%) 1 (1%)
10 (23%) 11 (25%) 9 (21%) 13 (29%) 1 (2%) 14 (32%) 0 (0%)
14 (29%) 6 (12%) 5 (10%) 22 (45%) 2 (4%) 21 (43%) 2 (4%)
⬍.0001
2 (4%)b 32 (73%) 25 5 2
7 (14%) 4 3
128 (100%) 71 (85%) 49 13 9 1 (1%)
⬍.0001
⬍.0001
3 (6%)
12 (14%) 6 6
10 (23%) 6 4
36 (74%) 22 14
NA NA
NA 10 ⫾ 11
41/3 13 ⫾ 14
3 (6%) 8/41 28 ⫾ 22
⬍.0001 ⬍.0001
NA
NA
3⫾3
8⫾8
⬍.0001
NA NA
NA NA
3 (7%) 2.6 (0.2–11)
17 (35%) 20.0 (7–70)
⬍.0001 ⬍.0001
3586 ⫾ 901
3599 ⫾ 977
3767 ⫾ 643
3803 ⫾ 650
.58
53 (41%) 43 (13– 83)
45 (54%) 48 (14 – 84)
31 (70%) 48 (23– 83)
47 (96%) 53 (14 – 84)
⬍.0001 .06
Abbreviations: C, central; F, female; L, lateral; LN, lymph node; M, male; NA, not applicable; THW, thyroid hormone withdrawal. a
Diffuse sclerosing variant, tall-cell, or columnar cell variant; oncocytic variant; or solid variant.
b
Microscopically involved nodes incidentally resected at time of thyroidectomy.
trast, 74% of pN1 macroscopic patients underwent central and lateral LND. In comparison with pN0 patients, pNx patients had smaller tumor sizes (P ⬍ .001), higher proportions of multifocal lesions (P ⬍ .001), and slightly lower proportions of extrathyroid extension (P ⬍ .05). A higher proportion of extrathyroid extension in pN1 microscopic than in pNx (P ⬍ .0001) or pN0 patients (P ⫽ .04) was the only prognosis-related difference between these subgroups. The pN1 microscopic and macroscopic patients
differed for the largest metastatic nodal focus [2.6 (range 0.2–11) vs 20.0 mm (range 7–70), P ⬍ .0001], the proportion of N1b (7% vs 84%, P ⬍ .0001), the number of metastatic lymph nodes (three vs eight, P ⬍ .0001), and the proportion of nodal ECE (7% vs 35%, P ⬍ .0001). Finally, as compared with pNx patients, pN1 macroscopic patients presented larger tumor sizes (P ⫽ .03) and higher proportions of aggressive pathological subtypes (P ⫽ .03), multifocal lesions (P ⫽ .03), and extrathyroid extension (P ⬍ .0001).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
136
Bardet et al
Microscopic Lymph Node Involvement
Risk of structural PRD and disease status at last follow-up During a median follow-up of 48 months (range 13– 84), 51 patients (17%) presented with structural persistent (n ⫽ 45) or recurrent (n ⫽ 6) disease (Supplemental Table 1). PRD was located only in the neck in 32 patients, only in distant sites in 11 patients, and in both the neck and distant sites in eight patients. Of those 51 patients, PRD was detected on postablation SPECT-CT in 32 (63%) and on FDG-PET-CT in 16 (31%) and was pathologically proven in 23 (45%). The stimulated Tg level at ablation was greater than 2 ng/mL in all TgAb negative patients but two with rhTSH-stimulated Tg level at 1.1 ng/mL (patients 1 and 8). Small-volume PRD was, however, clearly demonstrated in those two patients on postablation SPECT-CT images (abnormal lateral neck focus corresponding to a 4 mm lymph node in patient 1 and small lung foci without obvious nodules on corresponding computed tomography in patient 8). The cumulative probability of structural PRD was significantly different (P ⬍ .0001) among the four N subgroups (Figure 1). The 4-year cumulative risk of PRD was higher in pN1 macroscopic than in pN1 microscopic patients (49% vs 24%, P ⫽ .03) and higher in pN1 microscopic than in pN0 (24% vs 12%, P ⫽ .01) or pNx patients (24% vs 6%, P ⬍ .001). No significant difference was found between pN0 and pNx patients (P ⫽ .1). The proportions of large-volume lesions tended to differ between the four N subgroups (71% in pNx, 30% in pN0, 45% in pN1 microscopic, and 74% in pN1 macroscopic, P ⫽ .07). No difference was found for the proportions of RAI-refractory lesions (43% in pNx, 50% in pN0, 36% in pN1 microscopic, and 52% in pN1 macroscopic, P ⫽ .84). At the last visit, the proportion of patients with NED decreased significantly from pNx (98%), pN0 (93%), and pN1 microscopic (89%) compared with pN1 macroscopic patients (70%) (P ⬍ .0001, Cochran-Armitage trend test).
J Clin Endocrinol Metab, January 2015, 100(1):132–140
The proportion of patients with structural or biochemical evidence of persistent disease in the pNx, pN0, pN1 microscopic, or pN1 macroscopic subgroups were 1%, 5%, 4%, and 14%, and 1%, 2%, 7%, and 16%, respectively. Prognostic factor analysis Multivariate analysis controlling for age, gender, pathological subtypes, tumor size, multifocal tumors, bilateral tumors, extrathyroid extension, and pN status showed tumor size of 20 mm or greater [relative risk (RR) 3.4; P ⫽ .0001], extrathyroid extension (RR 2.6; P ⬍ .003), and pN1 macroscopic (RR 4.5; P ⬍ .0001) but also pN1 microscopic (RR 2.5; P ⬍ .02) to be independent risk factors for structural PRD (Table 2). Similar risk factors were found when focusing on locoregional disease, ie, tumor size of 20 mm or greater (RR 2.8; P ⬍ .01), extrathyroid extension (RR 2.5; P ⬍ .01), pN1 macroscopic (RR 7.1; P ⬍ .0001), and pN1 microscopic (RR 4.9; P ⬍ .001), and a bilateral thyroid tumor (RR 2.4; P ⬍ .01). Using similar variables, a multivariate analysis showed that extrathyroid extension [odds ratio (OR) 9.7; P ⬍ .0001] and pN1 macroscopic (OR 4.9; P ⬍ .001) were also independent predictors for structural or biochemical persistent disease at the last visit, but pN1 microscopic was not (Table 3). To assess the prognostic value of other variables related to the N status, four additional multivariate analyses were conducted on both the risk of structural PRD and the final outcome by replacing the microscopic/macroscopic N status by one of the following variables: 1) the size of the largest metastatic nodal focus (ⱕ2 mm, 2–10 mm, or ⬎ 10 mm); 2) the N status according to the TNM classification (N1a/N1b); 3) the number of metastatic lymph nodes (one to five or more than five); and 4) the presence or absence of lymph node ECE. The results of these final models using each of the above N-related variables adjusted for the other prognostic factors are presented in Tables 4 and 5. Although the largest metastatic nodal focus greater than 10 mm or between 2 and 10 mm predicted both structural PRD and persistent disease at the last follow-up visit, a size of 2 mm or less did not. Likewise, if both N1a and N1b predicted structural PRD, only N1b was associated with persistent disease at the last visit. Similar data were observed with the presence or absence of nodal ECE. Finally, both categories based on the number of lymph nodes predicted structural PRD and final status.
Discussion Figure 1. Cumulative probability of structural persistent/recurrent disease (locoregional and/or distant disease) according to the pathological pN status.
The present study in 305 consecutive PTC patients treated with surgery and RAI, and homogenously assessed by
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
doi: 10.1210/jc.2014-1199
Table 2.
jcem.endojournals.org
137
Results of Multivariate Analysis on Structural PRD: Initial and Final Models Initial Model
Variable Age, y ⬍45 ⱖ45 Gender Female Male Aggressive pathological subtypes No Yes Tumor size, mm ⬍20 ⱖ20 Multifocal tumor No Yes Bilateral tumor No Yes Extrathyroid extension No Yes pN status pNx pN0 pN1 microscopic pN1 macroscopic
Final Model
Patients at Risk, n
RR
95% CI
P Value
126 179
1.0 0.9
0.5–1.7
.81
238 67
1.0 1.3
0.7–2.4
.53
276 29
1.0 1.8
0.8 – 4.2
.16
170 135
1.0 3.8
2.0 –7.2
⬍.0001
160 145
1.0 1.7
0.7– 4.5
.26
197 108
1.0 1.0
0.4 –2.6
.96
249 56
1.0 2.7
1.4 –5.2
.003
128 84 44 49
1.0 1.8 4.0 6.8
0.7– 4.8 1.5–10.8 2.6 –17.3
.26 ⬍.02 ⬍.0001
RR
95% CI
P Value
1.0 3.4
1.8 – 6.4
.0001
1.0 2.6
1.4 – 4.7
⬍.003
1.2–5.5 2.3–9.1
⬍.02 ⬍.0001
1.0a 2.5 4.5
Abbreviation: CI, confidence interval. a Because the risk of PRD in the final model was not statistically different between pNx and pN0, both groups were pooled and set as the reference category.
postablation SPECT-CT scan, strongly suggests that PTC patients with initial microscopic lymph node involvement have an intermediate outcome between N0 –Nx patients and N1 macroscopic patients. Presently no consensus exists regarding the definition of microscopic and macroscopic nodal involvement. A classification based on the size of the largest nodal deposit with micrometastases (⬍2 mm), small nodal metastases (2–10 mm), intermediate (10 –30 mm), and large nodal metastases (⬎30 mm) was proposed (1) but is not yet used routinely. Although less accurate, the distinction between cN0 and cN1 recommended by the AJCC TNM staging system is simple and easy to use. In the present study, the size of the largest metastatic nodal focus was 0.2–11 mm in the N1 microscopic group and 7–70 mm in the macroscopic group, suggesting that, despite a slight overlap, 1 cm could be considered as an appropriate size cutoff between these two subgroups. Likewise, N1 macroscopic showed a higher number of metastatic lymph nodes, a higher proportion of ECE, and a higher proportion of lateral nodal involvement, all known factors associated with poor prognosis (15–18). These data confirm that the distinction between N1 macroscopic and N1 microscopic
is robust and valid and can delineate postoperatively patients with different prognosis. Although the follow-up was shorter in the present cohort than in our previous report (2), the proportion of patients with structural PRD was higher in the current study than in the previous one (17% vs 7%). This trend was observed for each N category, ie, pNx (6% vs 4%), pN0 (12% vs 2%), pN1 microscopic (24% vs 4%), and pN1 macroscopic (49% vs 19%). There are several explanations for these discrepancies. In the present study, we examined not only PRD in the neck (78% of our cases) but also distant disease, whereas the previous study focused on locoregional disease. Most importantly, the imaging methods we use routinely since 2006 are more sensitive than before. SPECT-CT particularly improves the interpretation of postablation total body scan detecting smallvolume iodine avid lesions (7, 8), thereby enabling the identification of a large proportion of patients with PRD. For RAI refractory patients, FDG-PET-CT was helpful by demonstrating lesions in about one-third of patients with PRD. Neck US was also a useful tool to confirm local PRD. Although US was performed preoperatively, most often by general radiologists focusing on thyroid nodules, it cannot
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
138
Bardet et al
Microscopic Lymph Node Involvement
J Clin Endocrinol Metab, January 2015, 100(1):132–140
Table 3. Results of Multivariate Analysis on Persistent Structural or Biochemical Disease at Last Follow-Up: Initial and Final Models Initial Model Variable Age, y ⬍45 ⱖ45 Gender Female Male Aggressive pathological subtypes No Yes Tumor size, mm ⬍20 ⱖ20 Multifocal tumor No Yes Bilateral tumor No Yes Extrathyroid extension No Yes pN status pNx pN0 pN1 microscopic pN1 macroscopic
Final Model
Patients at Risk, n
OR
95% CI
P Value
126 179
1.0 1.9
0.7–5.4
.24
238 67
1.0 1.3
0.4 –3.7
.66
276 29
1.0 1.2
0.3– 4.8
.76
170 135
1.0 1.5
0.6 – 4.1
.42
160 145
1.0 0.9
0.2– 4.5
.84
197 108
1.0 0.4
0.1–2.3
.32
249 56
1.0 6.3
2.4 –16.9
⬍.001
128 84 44 49
1.0 4.3 5.9 16.5
0.8 –24.0 1.0 –35.7 3.1– 88.2
.10 .05 .001
OR
95% CI
P Value
1.0 9.7
3.9 –24.1
⬍.0001
2.0 –12.2
⬍.001
1.0a 4.9
Abbreviation: CI, confidence interval. a Because the risk of persistent disease at last follow-up in the final model was not statistically different between pNx, pN0 and pN1 microscopic, all these groups were pooled and set as the reference category.
be excluded that nodal involvement could have been missed in a few patients falsely considered as cN0 patients. A multivariate analysis confirmed that the presence of macroscopic nodal involvement was the most powerful predictor (RR 4.5) of PRD but that microscopic nodal involvement was also identified as an independent risk factor of PRD (RR 2.5), in contrast to our first study (2). This difference is probably related to the SPECT-CT’s capacity to detect small-volume lesions in or outside the neck, as previously discussed. The present data show, however, that only small nodal metastases (2–10 mm), but not micrometastases (⬍2 mm), were found to predict PRD. A limited number of patients in each subgroup may be an explanation for those statistical results, even if there is a rationale for a correlation between the nodal size and the risk of PRD including the pN1 microscopic patients. Moreover, the multivariate analyses using the N1a/N1b classification, or the nodal ECE, provide other evidence for this relationship. Assuming that the proportion of microscopic nodal involvement is equivalent in the pNx and pN1 microscopic patients, because both are cN0, we could have expected a
similar rate of PRD for these two categories of patients. However, Nx patients had a better outcome than the N1 microscopic patients. This suggests a selection bias for LND and is in accordance with the lower proportion of extrathyroid extension in pNx than in pN1 microscopic patients (6% vs 32%), the only prognosis-related difference. Indeed, a multivariate analysis confirmed that extrathyroid extension remained a strong (RR 2.6) risk factor for PRD as previously shown (2, 15). In addition, tumor size larger than 20 mm, but not male gender, was found as a significant risk factor (RR 3.4) in accordance with earlier reports (5, 19). In the present study, we did not include the LND modalities into the multivariate model because we believe that the benefits of prophylactic LND on locoregional PRD can be assessed only in prospective randomized trials. Data also demonstrate that the proportion of patients with NED at the last visit decreased progressively from pNx-pN0 (98%–93%) and pN1 microscopic (89%) to pN1 macroscopic patients (70%). However, although macroscopic nodal involvement (OR 4.9) and extrathyroid extension (OR 9.7) were independent predictors for
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
doi: 10.1210/jc.2014-1199
jcem.endojournals.org
Table 4. Final Models of Multivariate Analysis on Structural PRD Using Four Variables Related to N Status Adjusted for Thyroid Tumor Size and Extrathyroid Extension Model
Patients at Risk, n
RR
95% CI
Table 5. Final Models of Multivariate Analysis on Persistent Structural or Biochemical Disease at Last Follow-Up Visit Using Four Variables Related to N Status Adjusted for Extrathyroid Extension
P Value
Model
A, using the size of the largest metastatic nodal focus pNx 128 1.0a pN0 84 ⱕ2 mm 19 2–10 mm 21 2.6 1.1– 6.3 ⬍.04 ⬎10 mm 49 3.8 2.0 –7.4 ⬍.0001 B, using the TNM classification pNx 128 1.0 pN0 84 N1a 49 3.0 1.4 – 6.4 ⬍.01 N1b 44 4.5 2.2–9.2 ⬍.0001 C, using the number of metastatic lymph nodes pNx 128 1.0 pN0 84 1–5 60 3.1 1.5– 6.4 ⬍.01 ⬎5 32 4.6 2.2–9.7 ⬍.0001 D, using the LN ECE pNx 128 1.0 pN0 84 No ECE 72 3.5 1.8 – 6.9 ⬍.001 ECE 20 4.2 1.8 –9.8 ⬍.01
Patients at Risk, n
OR
95% CI
Abbreviations: CI, confidence interval; LN, lymph node.
a
a
persistent disease, microscopic nodal involvement was not. This suggests that PRD associated with micrometastases was more easily curable than that associated with macroscopic nodal involvement. Interestingly, although the proportions of RAI refractory lesions were similar in pN1 microscopic or macroscopic patients, small-volume lesions tended to be more frequent in pN1 microscopic patients (55%) than in pN1 macroscopic patients (26%). Furthermore, treatment for PRD in pN1 macroscopic patients was intensified compared with pN1 microscopic patients, ie, surgery alone (35% vs 18%), RAI treatment alone (17% vs 36%), surgery and RAI treatment (26% vs 18%), and no additional treatment (13% vs 18%). Although the clinical significance of biochemical disease is variable, with some patients evolving toward a clinical recurrence and others showing a stable or slowly decreasing Tg level over time (20), the present data suggest a higher rate of residual disease at 4 years in pN1 microscopic and, even more, in pN1 macroscopic patients than in pNx-pN0 patients. This retrospective study presents several limitations, particularly due to the diversity of surgeons and surgical procedures. However, most patients were operated on in
P Value
A, using the size of the largest metastatic nodal focus pNx 128 1.0a pN0 84 ⱕ2 mm 19 2–10 mm 21 5.6 1.4 –22.5 ⬍.02 ⬎10 mm 49 5.5 2.0 –14.8 ⬍.001 B, using the TNM classification pNx 128 1.0 pN0 84 N1a 49 N1b 44 3.7 1.5–9.4 ⬍.01 C, using the number of metastatic lymph nodes pNx 128 1.0 pN0 84 1–5 60 3.2 1.1–9.0 ⬍.04 ⬎5 32 4.4 1.4 –14.3 ⬍.02 D, using the LN ECE pNx 128 1.0 pN0 84 No ECE 72 ECE 20 5.3 1.6 –17.2 ⬍.01
Abbreviations: CI, confidence interval; LN, lymph node. Because the risk of PRD was not statistically different between pNx, pN0, and 2 mm or less in model A and between pNx and pN0 in models B, C, and D, these groups were pooled and set as the reference category.
139
Because the risk of persistent disease at last follow-up visit was not statistically different between pNx, pN0, and 2 mm or less in model A, between pNx, pN0, and pN1a in model B, between pNx and pN0 in model C, and between pNx, pN0, and no ECE in model D, these groups were pooled and set as the reference category.
specialized institutions and most by head and neck surgeons who have shared similar training. Furthermore, the concept of microscopic/macroscopic LN involvement has been developed for several years and is generally adopted by the referring surgeons. The elaboration in 2009 of regional recommendations for thyroid cancer management also contributed to homogenize the indications and procedures of LND as well as the drafting of the surgical report. Therefore, we believe that we can confidently rely on the peroperative data of patients managed inside or outside our institution. The diversity of LND modalities could also account for differences in the risk of PRD observed in pN0 or pN1 patients. However, even if a similar proportion of pN0 or pN1 microscopic patients underwent a prophylactic LND, the risk of PRD was significantly higher in pN1 microscopic than in pN0 patients. Furthermore, even if pN1 macroscopic patients were submitted in general to a more extended LND than the pN1 microscopic patients, their risk of PRD remained significantly higher in comparison with pN1 microscopic patients. This reinforces our conclusion that pN0, pN1 microscopic, or pN1 macroscopic PTC patients present different levels of risk of PRD.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
140
Bardet et al
Microscopic Lymph Node Involvement
At a time when therapeutic deescalation is considered in PTC patients, our data suggest that recognizing microscopic nodal involvement in a cN0 patient remains an important issue. Despite potential complications, prophylactic central dissection is the only way to identify microscopic nodal involvement. Furthermore, post-operative RAI still may be indicated in N1 microscopic patients, at least in those with small nodal metastases (2–10 mm), for two reasons. First, post-RAI scintigraphy with SPECT-CT is likely to be the most reliable imaging modality to detect small iodine-avid residual lesions. Second, RAI treatment is the most appropriate tool to destroy such lesions (21). If randomized prospective studies are necessary to address the issue of the benefit of RAI in patients with or without microscopic nodal involvement, the diagnostic and therapeutic impacts of RAI might introduce some bias between the ablated and the nonablated group. In conclusion, the present study demonstrates that the risk of PRD is likely higher than expected with the use of sensitive imaging tools and that patients with initial microscopic lymph node involvement present an intermediate outcome between that observed in N0-Nx patients and N1 macroscopic patients. Although long-term follow-up is needed, these retrospective data provide a new basis for discussing the management of PTC patients and could be taken into account for modifications to the current risk recurrence staging systems.
J Clin Endocrinol Metab, January 2015, 100(1):132–140
4.
5. 6.
7.
8.
9.
10.
11.
12.
13.
14.
Acknowledgments We greatly thank Professor Martin Schlumberger (Institut Gustave Roussy, Villejuif, France) for his helpful discussion. We are indebted to Helen Lapasset for her assistance in reviewing the manuscript. Address all correspondence and requests for reprints to: Stéphane Bardet, MD, Department of Nuclear Medicine and Thyroid Unit, Centre François Baclesse, 3 Avenue Général Harris, BP 5026, F-14076 Caen, Cedex 05, France. E-mail: [email protected]. Disclosure Summary: The authors have nothing to declare.
15.
16.
17.
18.
19.
References 1. Randolph GW, Duh QY, Heller KS, et al. The prognostic significance of nodal metastases from papillary thyroid carcinoma can be stratified based on the size and number of metastatic lymph nodes, as well as the presence of extranodal extension. Thyroid. 2012;22: 1144 –1152. 2. Bardet S, Malville E, Rame JP, et al. Macroscopic lymph-node involvement and neck dissection predict lymph-node recurrence in papillary thyroid carcinoma. Eur J Endocrinol. 2008;158:551–560. 3. Ito Y, Tomoda C, Uruno T, et al. Ultrasonographically and anatomopathologically detectable node metastases in the lateral compart-
20.
21.
ment as indicators of worse relapse-free survival in patients with papillary thyroid carcinoma. World J Surg. 2005;29:917–920. Wada N, Duh QY, Sugino K, et al. Lymph node metastasis from 259 papillary thyroid microcarcinomas: frequency, pattern of occurrence and recurrence, and optimal strategy for neck dissection. Ann Surg. 2003;237:399 – 407. Edge SB, Byrd DR, Compton CC, Fritz AG, Green FL, Trotti A. AJCC Cancer Staging Manual. New York: Springer; 2010. Cooper DS, Doherty GM, Haugen BR, et al. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid. 2009;19:1167–1214. Aide N, Heutte N, Rame JP, et al. Clinical relevance of single-photon emission computed tomography/computed tomography of the neck and thorax in postablation (131)I scintigraphy for thyroid cancer. J Clin Endocrinol Metab. 2009;94:2075–2084. Grewal RK, Tuttle RM, Fox J, et al. The effect of posttherapy 131I SPECT/CT on risk classification and management of patients with differentiated thyroid cancer. J Nucl Med. 2010;51:1361–1367. Schmidt D, Szikszai A, Linke R, Bautz W, Kuwert T. Impact of 131I SPECT/spiral CT on nodal staging of differentiated thyroid carcinoma at the first radioablation. J Nucl Med. 2009;50:18 –23. Leboulleux S, El Bez I, Borget I, et al. Post-radioiodine treatment whole body scan in the era of fluorodesoxyglucose positron emission tomography for differentiated thyroid carcinoma with elevated serum thyroglobulin levels. Thyroid. 2012;22:832– 838. De Lellis RA, Lloyd RV, Heitz PU, Eng C, eds. Word Health Organization Classification of Tumours. Pathology and Genetics of Tumours of Endocrine Organs. Lyon, France: IARC Press; 2004. Schlumberger M, Catargi B, Borget I, et al. Strategies of radioiodine ablation in patients with low-risk thyroid cancer. N Engl J Med. 2012;366:1663–1673. Tuttle RM, Tala H, Shah J, et al. Estimating risk of recurrence in differentiated thyroid cancer after total thyroidectomy and radioactive iodine remnant ablation: using response to therapy variables to modify the initial risk estimates predicted by the new American Thyroid Association staging system. Thyroid. 2010;20:1341–1349. Schlumberger M, Sherman SI. Approach to the patient with advanced differentiated thyroid cancer. Eur J Endocrinol. 2012;166:5–11. Leboulleux S, Rubino C, Baudin E, et al. Prognostic factors for persistent or recurrent disease of papillary thyroid carcinoma with neck lymph node metastases and/or tumor extension beyond the thyroid capsule at initial diagnosis. J Clin Endocrinol Metab. 2005; 90:5723–5729. Machens A, Dralle H. Correlation between the number of lymph node metastases and lung metastasis in papillary thyroid cancer. J Clin Endocrinol Metab. 2012;97:4375– 4382. Vas Nunes JH, Clark JR, Gao K, et al. Prognostic implications of lymph node yield and lymph node ratio in papillary thyroid carcinoma. Thyroid. 2013;23:811– 816. Verburg FA, Mader U, Tanase K, et al. Life expectancy is reduced in differentiated thyroid cancer patients ⱖ 45 years old with extensive local tumor invasion, lateral lymph node, or distant metastases at diagnosis and normal in all other DTC patients. J Clin Endocrinol Metab. 2013;98:172–180. Hay ID, Bergstralh EJ, Goellner JR, Ebersold JR, Grant CS. Predicting outcome in papillary thyroid carcinoma: development of a reliable prognostic scoring system in a cohort of 1779 patients surgically treated at one institution during 1940 through 1989. Surgery. 1993;114:1050 –1057. Baudin E, Do CC, Cailleux AF, Leboulleux S, Travagli JP, Schlumberger M. Positive predictive value of serum thyroglobulin levels, measured during the first year of follow-up after thyroid hormone withdrawal, in thyroid cancer patients. J Clin Endocrinol Metab. 2003;88:1107–1111. Schmidt D, Linke R, Uder M, Kuwert T. Five months’ follow-up of patients with and without iodine-positive lymph node metastases of thyroid carcinoma as disclosed by (131)I-SPECT/CT at the first radioablation. Eur J Nucl Med Mol Imaging. 2010;37:699 –705.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
ARTICLE C a r e
Prognostic Value of Microscopic Lymph Node Involvement in Patients With Papillary Thyroid Cancer Stéphane Bardet, Renaud Ciappuccini, Elske Quak, Jean-Pierre Rame, David Blanchard, Dominique de Raucourt, Emmanuel Babin, Jean-Jacques Michels, Dominique Vaur, and Natacha Heutte Department of Nuclear Medicine and Thyroid Unit (S.B., R.C., E.Q.), and Departments of Head and Neck Surgery (J.-P.R., D.B., D.d.R.), and Pathology (J.-J.M.), Biology (D.V.), Centre François Baclesse, 14076 Caen, France; Department of Head and Neck Surgery (E.B.), University Hospital, Caen 14000, France; Unité 1086 (N.H.), INSERM-University of Caen-Basse Normandie, “Cancers and Préventions” Program, University of Caen-Basse Normandie, 14032 Caen, France
Context: The impact of microscopic nodal involvement on the risk of persistent/recurrent disease (PRD) remains controversial in patients with papillary thyroid carcinoma (PTC). Objective: The goal of the study was to assess the risk of PRD and the 4-year outcome in PTC patients according to their initial nodal status [pNx, pN0, pN1 microscopic (cN0/pN1) or pN1 macroscopic (cN1/pN1)]. Design: We conducted a retrospective cohort study. Patients: The study included 305 consecutive PTC patients referred for radioiodine ablation from 2006 to 2011. Main Outcome Measure: We evaluated the risk of structural PRD and the disease status at the last follow-up. At ablation, persistent disease was consistently assessed by using post-radioiodine ablation scintigraphy combining total body scan and neck and thorax single-photon computed tomography-computed tomography (SPECT-CT) acquisition. Results: Of 305 patients, 128 (42%) were pNx, 84 (28%) pN0, 44 (14%) pN1 microscopic, and 49 (16%) pN1 macroscopic. The 4-year cumulative risk of PRD was higher in pN1 macroscopic than in pN1 microscopic patients (49% vs 24%, P ⫽ .03), and higher in pN1 microscopic than in pN0 (12%, P ⫽ .01) or pNx patients (6%, P ⬍ .001). On multivariate analysis, tumor size of 20 mm or greater [relative risk (RR) 3.4; P ⫽ .0001], extrathyroid extension (RR 2.6; P ⬍ .003), pN1 macroscopic (RR 4.5; P ⬍ .0001), and pN1 microscopic (RR 2.5; P ⬍ .02) were independent risk factors for PRD. At the last visit, the proportion of patients with no evidence of disease decreased from pNx (98%), pN0 (93%), and pN1 microscopic (89%) to pN1 macroscopic patients (70%) (P ⬍ .0001, Cochran-Armitage trend test). Extrathyroid extension (odds ratio 9.7; P ⬍ .0001) and N1 macroscopic (OR 4.9; P ⬍ .001) independently predicted persistent disease at the last visit, but N1 microscopic did not. Conclusions: Patients with microscopic lymph node involvement present an intermediate outcome between that observed in pN0-pNx patients and pN1 macroscopic patients. These data may justify modifications to the risk recurrence staging systems. (J Clin Endocrinol Metab 100: 132–140, 2015)
ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2015 by the Endocrine Society Received January 16, 2014. Accepted September 30, 2014. First Published Online October 10, 2014
132
jcem.endojournals.org
Abbreviations: AJCC, American Joint Committee on Cancer; ECE, extracapsular extension; FDG-PET-CT, 18F-fluorodeoxyglucose-positron emission tomography-computed tomography; LND, lymph node dissection; NED, no evidence of disease; OR, odds ratio; PRD, persistent/recurrent disease; PTC, papillary thyroid carcinoma; RAI, radioiodine; rhTSH, recombinant human TSH RR, relative risk; SPECT-CT, single-photon emission computed tomography-computed tomography; Tg, thyroglobulin; TgAb, Tg antibody; TNM, tumor node metastasis; US, ultrasound.
J Clin Endocrinol Metab, January 2015, 100(1):132–140
doi: 10.1210/jc.2014-1199
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
doi: 10.1210/jc.2014-1199
he presence of nodal metastases on the primary pathological report (pN1) is known to increase the risk of locoregional persistent/recurrent disease (PRD) and, in some instances, the risk of disease-specific mortality in patients with papillary thyroid carcinoma (PTC). The magnitude of risk depends, however, on several characteristics of the nodal involvement, mainly its volume (1). Schematically, large-volume, macroscopic, clinically apparent nodal involvement has been associated with rates of recurrence of 20%–25% (2– 4). Conversely, small-volume, microscopic nodal involvement, has been related to rates below 5% (2, 4), not clearly higher than those estimated in patients without nodal involvement (pN0) after prophylactic lymph node dissection (LND) or unknown nodal status (pNx). However, the presence of nodal involvement, regardless of size, has an upstaging impact in both the American Joint Committee on Cancer (AJCC) tumor node metastasis (TNM) staging system (5) and the American Thyroid Association risk of recurrence system (6). In a previous study on 545 PTC patients diagnosed between 1965 and 2003, of whom 91% have had radioiodine (RAI), the 10-year cumulative risk of lymph node recurrence was estimated at 7% with rates at 4% in pNx, 2% in pN0, 4% in pN1 microscopic, and 19% in pN1 macroscopic patients (2). Macroscopic, but not microscopic, lymph node involvement was found as an independent risk factor of lymph node recurrence as well as extrathyroidal invasion and male gender. In the last 10 years, in addition to ultrasound (US), new functional imaging modalities have emerged improving the detection of small (⬍10 mm) lesions, in or outside the neck, which can be either not visible or considered as benign on conventional radiological modalities. These include single-photon emission computed tomography-computed tomography (SPECT-CT) imaging during 131I scintigraphy, mainly in the postablation setting (7–9), and 18F-fluorodeoxyglucose-positron emission tomography-computed tomography (FDG-PET-CT) scan in the context of high serum thyroglobulin (Tg) levels, especially in iodine-refractory patients (10). The aim of the present study was therefore to reassess the risk of structural persistent/recurrent locoregional and distant disease in newly diagnosed PTC patients since 2006, treated with surgery and postoperative RAI, and to reexamine the impact of microscopic or macroscopic lymph node involvement on patient outcome.
T
jcem.endojournals.org
133
and subsequently followed up at our institute, were initially included in the present study. Three patients with a poorly differentiated component were excluded and 305 patients were retained for final analysis.
Surgery Surgery was performed either at our institution or at the university hospital (n ⫽ 199) or in other institutions (n ⫽ 106). Most patients (n ⫽ 264) were operated on by head and neck surgeons. Preoperative neck US for the assessment of thyroid nodules was performed by radiologists working inside or outside our institution. All patients underwent a total thyroidectomy in one (n ⫽ 233) or two interventions (n ⫽ 72). The decision to perform or not to perform a prophylactic or a therapeutic LND, and its extent, depended on several factors including the cN0/cN1 distinction of the AJCC TNM staging system (5), the surgeon’s preference or experience, and the preoperative cytological and the frozen section examination data when performed. In cN0 patients, surgery included either a prophylactic LND of the central compartment (level VI) sometimes associated with LND of the lateral compartments (levels IIa, III, and IV), or no LND at all. Central LND was bilateral in most cases, could be ipsilateral to the thyroid tumor only when frozen section examination was dubious, and rarely only contralateral when it was falsely negative, leading to total thyroidectomy in two interventions. In cN1 patients, a therapeutic LND was performed generally consisting of a central and lateral LND. The criteria applied for defining initial macroscopic or microscopic lymph node involvement have been previously reported (2). A lymph node metastasis was considered macroscopic (cN1/pN1 for the AJCC) when evident on clinical examination or US before surgery and/or clearly suspected by the surgeon according to his report. All other lymph node metastases were considered microscopic (cN0/pN1).
Pathology Diagnosis of PTC was based on the written pathological report. Pathological variants were defined according to the World Health Organization classification (11). Based on the pathological report, the number of resected and metastatic lymph nodes and the presence of nodal extracapsular extension (ECE) were notified for each patient. When available, the slides were reviewed to estimate the size of the largest metastatic nodal focus.
Radioiodine ablation RAI was administered either after thyroid hormone withdrawal or after recombinant human TSH (rhTSH) (0.9 mg, im, for 2 consecutive days; Thyrogen; Genzyme). The RAI activity (1.1 or 3.7 GBq) and the preparation modalities were determined after local multidisciplinary staff except for the 42 low-risk PTC patients randomized in the trial (12). A post-RAI total body scan was performed 2 or 5 days after 1.1 or 3.7 GBq, respectively, with SPECT-CT acquisition in all patients, at least on the neck and thorax, as previously described (7).
Tg and Tg antibody (TgAb) assay
Patients and Methods Patients Three hundred eight consecutive PTC patients referred to our department for RAI between January 2006 and December 2011,
Blood samples for stimulated serum Tg and TgAb measurements were collected immediately before RAI administration, and stored at ⫺20°C. All serum Tg levels were measured using Roche Cobas 6000 Tg (Roche Diagnostics) with a detection limit of 0.1 ng/mL and a functional sensitivity of 1.0 ng/mL. TgAb was
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
134
Bardet et al
Microscopic Lymph Node Involvement
measured using Immulite TgAb (Diagnostic Product Corp) from January 2006 to June 2008 and thereafter using Roche Cobas 6000 anti-Tg assay (Roche Diagnostics).
Definition of structural persistent/recurrent disease As previously proposed (13), structural PRD was defined as evidence of tumor in the thyroid bed, lymph node metastases or distant metastases after completion of initial treatment (ie, surgery and RAI) confirmed by either histology or complementary imaging modalities. At ablation, persistent disease was consistently assessed using the post-RAI scintigraphy combining total body scan and at least neck and thorax SPECT-CT. In the presence of scintigraphic abnormalities, other imaging modalities such as neck US, neck and thorax CT scan with contrast medium injection, FDG-PET-CT, or bone magnetic resonance imaging was performed at the discretion of the clinician to provide additional evidence of residual disease and to better define the disease extent. For patients with normal postablation scintigraphy but persistent detectable serum Tg level (⬎1 ng/mL), similar imaging modalities were performed to evidence RAI refractory lesions. At that point, we chose not to consider biochemical disease, ie, elevated serum Tg level (suppressed Tg ⱖ 1 ng/mL and/or stimulated Tg ⱖ 2 ng/mL) with no other evidence of disease. We arbitrarily defined persistent disease when the diagnosis was made within 1 year after the initial treatment and recurrent disease after this period. To assess tumor burden, PRD was classified into two categories. Small-volume disease was defined either by the presence of clearly abnormal foci on post-RAI scintigraphy or on FDGPET-CT without obvious lesions on conventional radiology or by small (⬍10 mm) but highly suspicious lesions on conventional radiology (neck US, CT scan, magnetic resonance imaging). Conversely, large-volume disease was defined by large (⬎10 mm) lesions clearly evidenced by conventional radiology, whatever the presence of scintigraphic abnormalities. The RAI refractory feature of the PRD was also specified as previously defined (14).
Clinical outcome assessment As previously described (7, 12), patients with a normal postablation scan were seen at 3 months on levothyroxine (LT4) treatment for serum TSH, Tg, and TgAb determinations. When suppressed Tg level was less than 1 ng/mL in the absence of TgAb, disease status was assessed at 9 months by serum rhTSH-stimulated Tg determination and neck US. In the presence of TgAb, a diagnostic 131I scintigraphy after rhTSH was also performed. When evaluation at 9 months was normal (ie, stimulated Tg ⱕ 1 ng/mL and normal US or post-131I scintigraphy), patients were annually followed up with suppressed serum Tg and TgAb measurements. For patients with abnormal postablation scintigraphy, the clinical management depended on the location and size of persistent disease. In patients with nodal disease, surgery was favored for large-volume lesions, whereas close surveillance or additional RAI treatment was preferred for small-volume disease. For distant metastases, RAI treatment was generally repeated at 6-month intervals until the disappearance of abnormal foci. Neck irradiation was considered in patients with inoperable local extension. For patients with RAI refractory lesions, close surveillance or therapeutic options were discussed.
J Clin Endocrinol Metab, January 2015, 100(1):132–140
At the last follow-up visit, the disease status was categorized into one of the following three subgroups (13): 1) no evidence of disease (NED) defined by a suppressed serum Tg level less than 1 ng/mL and, if performed, normal imaging; 2) persistent structural disease defined by a Tg level of 1 ng/mL or greater associated with imaging abnormalities; and 3) persistent biochemical disease with a Tg level of 1 ng/mL or greater and no other evidence of disease. In the presence of TgAb, disease status assessment was based on imaging data and serum Tg if there was a level of 1 ng/mL or greater.
Data analysis Patient characteristics were compared using the 2 test, the Fisher’s exact test, or the Kruskal-Wallis test, as appropriate. The trend test of Cochran-Armitage was used to examine proportions of disease-related events according to N status in the following subgroups: Nx, N0, N1 microscopic, and N1 macroscopic. Time to PRD was calculated from the day of postablation scintigraphy to the date of diagnosis of PRD or February first 2013, whichever came first. The date of diagnosis was either the date of biopsy or surgery or that of abnormal imaging modalities or that of postablation scintigraphy when the persistent tumor was demonstrated on scintigraphy only. In this latter case, time to persistent disease was arbitrarily defined as 1 month. The Kaplan-Meier method was used to estimate the probability of PRD. The estimates in each group were compared using the logrank test. The analysis of prognostic factors was performed using the multivariate Cox regression model for PRD and using logistic regression with stepwise selection for persistent structural or biochemical disease at the end of follow-up. Statistical significance was defined as P ⬍ .05. All tests were two sided. SAS 9.3 statistical software (SAS Institute) was used to analyze the data.
Results Patient characteristics The study group included 305 patients with PTC (238 women and 67 men, mean age 49 ⫾ 17 y). There were 185 patients with a conventional form of PTC (61%), 91 with a follicular variant (30%), five with a tall-cell or a columnar-cell variant (2%), seven with a diffuse sclerosing variant (2%), three with an oncocytic variant (1%), and 14 with a solid variant (4%). After surgery, 128 patients (42%) were considered pNx, 84 (28%) pN0, 44 (14%) pN1 microscopic, and 49 (16%) pN1 macroscopic (Table 1). The four N subgroups did not differ for age, sex ratio, aggressive pathological subtypes, bilateral tumors, RAI activity, and median follow-up but differed for tumor size (P ⬍ .001), multifocal tumors (P ⬍ .01), T status of the TNM classification (P ⬍ .0001), and extrathyroid extension (P ⬍ .0001). The LND modalities did not significantly differ between pN0 and pN1 microscopic patients (P ⫽ .09) with prophylactic central LND performed in 85% vs 73% and central and lateral LND in 14% vs 23% of patients, respectively. In con-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
doi: 10.1210/jc.2014-1199
Table 1.
jcem.endojournals.org
135
Characteristics of Patients According to Initial pN Status
Mean age ⫾SD, y Sex ratio (F/M) Aggressive pathological subtypesa Mean tumor size ⫾ SD, mm Multifocal tumor Bilateral tumor T status (TNM) T1a T1b T2 T3 T4 Extrathyroid extension Known distant metastases before surgery Dissection modalities No dissection Central dissection only Bilateral Ipsilateral Contralateral Lateral dissection only (ipsilateral) C and L dissection Bilateral C and L Bilateral C and Ipsilateral L Adenectomy N1a/N1b Mean number of resected LN ⫾ SD Mean number of metastatic LN ⫾ SD LN ECE Median size of the largest metastatic LN, mm Mean I131 activity ⫾ SD, MBq Preparation with THW Median follow-up, mo
pNx (n ⴝ 128)
pN0 (n ⴝ 84)
pN1 Microscopic (n ⴝ 44)
pN1 Macroscopic (n ⴝ 49)
P Value
52 ⫾ 14 3.6 (101/27) 7 (5%)
48 ⫾ 16 4.2 (68/16) 8 (10%)
47 ⫾ 18 4.0 (36/8) 5 (11%)
47 ⫾ 21 1.9 (33/16) 9 (18%)
.18 .25 .07
17 ⫾ 13
23 ⫾ 10
19 ⫾ 12
25 ⫾ 24
⬍.001
75 (59%) 51 (40%)
29 (35%) 25 (30%)
21 (48%) 16 (36%)
20 (41%) 16 (33%)
⬍.01 .48
48 (38%) 39 (31%) 30 (23%) 10 (8%) 1 (1%) 8 (6%) 1 (1%)
9 (11%) 28 (33%) 31 (37%) 14 (17%) 2 (2%) 13 (15%) 1 (1%)
10 (23%) 11 (25%) 9 (21%) 13 (29%) 1 (2%) 14 (32%) 0 (0%)
14 (29%) 6 (12%) 5 (10%) 22 (45%) 2 (4%) 21 (43%) 2 (4%)
⬍.0001
2 (4%)b 32 (73%) 25 5 2
7 (14%) 4 3
128 (100%) 71 (85%) 49 13 9 1 (1%)
⬍.0001
⬍.0001
3 (6%)
12 (14%) 6 6
10 (23%) 6 4
36 (74%) 22 14
NA NA
NA 10 ⫾ 11
41/3 13 ⫾ 14
3 (6%) 8/41 28 ⫾ 22
⬍.0001 ⬍.0001
NA
NA
3⫾3
8⫾8
⬍.0001
NA NA
NA NA
3 (7%) 2.6 (0.2–11)
17 (35%) 20.0 (7–70)
⬍.0001 ⬍.0001
3586 ⫾ 901
3599 ⫾ 977
3767 ⫾ 643
3803 ⫾ 650
.58
53 (41%) 43 (13– 83)
45 (54%) 48 (14 – 84)
31 (70%) 48 (23– 83)
47 (96%) 53 (14 – 84)
⬍.0001 .06
Abbreviations: C, central; F, female; L, lateral; LN, lymph node; M, male; NA, not applicable; THW, thyroid hormone withdrawal. a
Diffuse sclerosing variant, tall-cell, or columnar cell variant; oncocytic variant; or solid variant.
b
Microscopically involved nodes incidentally resected at time of thyroidectomy.
trast, 74% of pN1 macroscopic patients underwent central and lateral LND. In comparison with pN0 patients, pNx patients had smaller tumor sizes (P ⬍ .001), higher proportions of multifocal lesions (P ⬍ .001), and slightly lower proportions of extrathyroid extension (P ⬍ .05). A higher proportion of extrathyroid extension in pN1 microscopic than in pNx (P ⬍ .0001) or pN0 patients (P ⫽ .04) was the only prognosis-related difference between these subgroups. The pN1 microscopic and macroscopic patients
differed for the largest metastatic nodal focus [2.6 (range 0.2–11) vs 20.0 mm (range 7–70), P ⬍ .0001], the proportion of N1b (7% vs 84%, P ⬍ .0001), the number of metastatic lymph nodes (three vs eight, P ⬍ .0001), and the proportion of nodal ECE (7% vs 35%, P ⬍ .0001). Finally, as compared with pNx patients, pN1 macroscopic patients presented larger tumor sizes (P ⫽ .03) and higher proportions of aggressive pathological subtypes (P ⫽ .03), multifocal lesions (P ⫽ .03), and extrathyroid extension (P ⬍ .0001).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
136
Bardet et al
Microscopic Lymph Node Involvement
Risk of structural PRD and disease status at last follow-up During a median follow-up of 48 months (range 13– 84), 51 patients (17%) presented with structural persistent (n ⫽ 45) or recurrent (n ⫽ 6) disease (Supplemental Table 1). PRD was located only in the neck in 32 patients, only in distant sites in 11 patients, and in both the neck and distant sites in eight patients. Of those 51 patients, PRD was detected on postablation SPECT-CT in 32 (63%) and on FDG-PET-CT in 16 (31%) and was pathologically proven in 23 (45%). The stimulated Tg level at ablation was greater than 2 ng/mL in all TgAb negative patients but two with rhTSH-stimulated Tg level at 1.1 ng/mL (patients 1 and 8). Small-volume PRD was, however, clearly demonstrated in those two patients on postablation SPECT-CT images (abnormal lateral neck focus corresponding to a 4 mm lymph node in patient 1 and small lung foci without obvious nodules on corresponding computed tomography in patient 8). The cumulative probability of structural PRD was significantly different (P ⬍ .0001) among the four N subgroups (Figure 1). The 4-year cumulative risk of PRD was higher in pN1 macroscopic than in pN1 microscopic patients (49% vs 24%, P ⫽ .03) and higher in pN1 microscopic than in pN0 (24% vs 12%, P ⫽ .01) or pNx patients (24% vs 6%, P ⬍ .001). No significant difference was found between pN0 and pNx patients (P ⫽ .1). The proportions of large-volume lesions tended to differ between the four N subgroups (71% in pNx, 30% in pN0, 45% in pN1 microscopic, and 74% in pN1 macroscopic, P ⫽ .07). No difference was found for the proportions of RAI-refractory lesions (43% in pNx, 50% in pN0, 36% in pN1 microscopic, and 52% in pN1 macroscopic, P ⫽ .84). At the last visit, the proportion of patients with NED decreased significantly from pNx (98%), pN0 (93%), and pN1 microscopic (89%) compared with pN1 macroscopic patients (70%) (P ⬍ .0001, Cochran-Armitage trend test).
J Clin Endocrinol Metab, January 2015, 100(1):132–140
The proportion of patients with structural or biochemical evidence of persistent disease in the pNx, pN0, pN1 microscopic, or pN1 macroscopic subgroups were 1%, 5%, 4%, and 14%, and 1%, 2%, 7%, and 16%, respectively. Prognostic factor analysis Multivariate analysis controlling for age, gender, pathological subtypes, tumor size, multifocal tumors, bilateral tumors, extrathyroid extension, and pN status showed tumor size of 20 mm or greater [relative risk (RR) 3.4; P ⫽ .0001], extrathyroid extension (RR 2.6; P ⬍ .003), and pN1 macroscopic (RR 4.5; P ⬍ .0001) but also pN1 microscopic (RR 2.5; P ⬍ .02) to be independent risk factors for structural PRD (Table 2). Similar risk factors were found when focusing on locoregional disease, ie, tumor size of 20 mm or greater (RR 2.8; P ⬍ .01), extrathyroid extension (RR 2.5; P ⬍ .01), pN1 macroscopic (RR 7.1; P ⬍ .0001), and pN1 microscopic (RR 4.9; P ⬍ .001), and a bilateral thyroid tumor (RR 2.4; P ⬍ .01). Using similar variables, a multivariate analysis showed that extrathyroid extension [odds ratio (OR) 9.7; P ⬍ .0001] and pN1 macroscopic (OR 4.9; P ⬍ .001) were also independent predictors for structural or biochemical persistent disease at the last visit, but pN1 microscopic was not (Table 3). To assess the prognostic value of other variables related to the N status, four additional multivariate analyses were conducted on both the risk of structural PRD and the final outcome by replacing the microscopic/macroscopic N status by one of the following variables: 1) the size of the largest metastatic nodal focus (ⱕ2 mm, 2–10 mm, or ⬎ 10 mm); 2) the N status according to the TNM classification (N1a/N1b); 3) the number of metastatic lymph nodes (one to five or more than five); and 4) the presence or absence of lymph node ECE. The results of these final models using each of the above N-related variables adjusted for the other prognostic factors are presented in Tables 4 and 5. Although the largest metastatic nodal focus greater than 10 mm or between 2 and 10 mm predicted both structural PRD and persistent disease at the last follow-up visit, a size of 2 mm or less did not. Likewise, if both N1a and N1b predicted structural PRD, only N1b was associated with persistent disease at the last visit. Similar data were observed with the presence or absence of nodal ECE. Finally, both categories based on the number of lymph nodes predicted structural PRD and final status.
Discussion Figure 1. Cumulative probability of structural persistent/recurrent disease (locoregional and/or distant disease) according to the pathological pN status.
The present study in 305 consecutive PTC patients treated with surgery and RAI, and homogenously assessed by
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
doi: 10.1210/jc.2014-1199
Table 2.
jcem.endojournals.org
137
Results of Multivariate Analysis on Structural PRD: Initial and Final Models Initial Model
Variable Age, y ⬍45 ⱖ45 Gender Female Male Aggressive pathological subtypes No Yes Tumor size, mm ⬍20 ⱖ20 Multifocal tumor No Yes Bilateral tumor No Yes Extrathyroid extension No Yes pN status pNx pN0 pN1 microscopic pN1 macroscopic
Final Model
Patients at Risk, n
RR
95% CI
P Value
126 179
1.0 0.9
0.5–1.7
.81
238 67
1.0 1.3
0.7–2.4
.53
276 29
1.0 1.8
0.8 – 4.2
.16
170 135
1.0 3.8
2.0 –7.2
⬍.0001
160 145
1.0 1.7
0.7– 4.5
.26
197 108
1.0 1.0
0.4 –2.6
.96
249 56
1.0 2.7
1.4 –5.2
.003
128 84 44 49
1.0 1.8 4.0 6.8
0.7– 4.8 1.5–10.8 2.6 –17.3
.26 ⬍.02 ⬍.0001
RR
95% CI
P Value
1.0 3.4
1.8 – 6.4
.0001
1.0 2.6
1.4 – 4.7
⬍.003
1.2–5.5 2.3–9.1
⬍.02 ⬍.0001
1.0a 2.5 4.5
Abbreviation: CI, confidence interval. a Because the risk of PRD in the final model was not statistically different between pNx and pN0, both groups were pooled and set as the reference category.
postablation SPECT-CT scan, strongly suggests that PTC patients with initial microscopic lymph node involvement have an intermediate outcome between N0 –Nx patients and N1 macroscopic patients. Presently no consensus exists regarding the definition of microscopic and macroscopic nodal involvement. A classification based on the size of the largest nodal deposit with micrometastases (⬍2 mm), small nodal metastases (2–10 mm), intermediate (10 –30 mm), and large nodal metastases (⬎30 mm) was proposed (1) but is not yet used routinely. Although less accurate, the distinction between cN0 and cN1 recommended by the AJCC TNM staging system is simple and easy to use. In the present study, the size of the largest metastatic nodal focus was 0.2–11 mm in the N1 microscopic group and 7–70 mm in the macroscopic group, suggesting that, despite a slight overlap, 1 cm could be considered as an appropriate size cutoff between these two subgroups. Likewise, N1 macroscopic showed a higher number of metastatic lymph nodes, a higher proportion of ECE, and a higher proportion of lateral nodal involvement, all known factors associated with poor prognosis (15–18). These data confirm that the distinction between N1 macroscopic and N1 microscopic
is robust and valid and can delineate postoperatively patients with different prognosis. Although the follow-up was shorter in the present cohort than in our previous report (2), the proportion of patients with structural PRD was higher in the current study than in the previous one (17% vs 7%). This trend was observed for each N category, ie, pNx (6% vs 4%), pN0 (12% vs 2%), pN1 microscopic (24% vs 4%), and pN1 macroscopic (49% vs 19%). There are several explanations for these discrepancies. In the present study, we examined not only PRD in the neck (78% of our cases) but also distant disease, whereas the previous study focused on locoregional disease. Most importantly, the imaging methods we use routinely since 2006 are more sensitive than before. SPECT-CT particularly improves the interpretation of postablation total body scan detecting smallvolume iodine avid lesions (7, 8), thereby enabling the identification of a large proportion of patients with PRD. For RAI refractory patients, FDG-PET-CT was helpful by demonstrating lesions in about one-third of patients with PRD. Neck US was also a useful tool to confirm local PRD. Although US was performed preoperatively, most often by general radiologists focusing on thyroid nodules, it cannot
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
138
Bardet et al
Microscopic Lymph Node Involvement
J Clin Endocrinol Metab, January 2015, 100(1):132–140
Table 3. Results of Multivariate Analysis on Persistent Structural or Biochemical Disease at Last Follow-Up: Initial and Final Models Initial Model Variable Age, y ⬍45 ⱖ45 Gender Female Male Aggressive pathological subtypes No Yes Tumor size, mm ⬍20 ⱖ20 Multifocal tumor No Yes Bilateral tumor No Yes Extrathyroid extension No Yes pN status pNx pN0 pN1 microscopic pN1 macroscopic
Final Model
Patients at Risk, n
OR
95% CI
P Value
126 179
1.0 1.9
0.7–5.4
.24
238 67
1.0 1.3
0.4 –3.7
.66
276 29
1.0 1.2
0.3– 4.8
.76
170 135
1.0 1.5
0.6 – 4.1
.42
160 145
1.0 0.9
0.2– 4.5
.84
197 108
1.0 0.4
0.1–2.3
.32
249 56
1.0 6.3
2.4 –16.9
⬍.001
128 84 44 49
1.0 4.3 5.9 16.5
0.8 –24.0 1.0 –35.7 3.1– 88.2
.10 .05 .001
OR
95% CI
P Value
1.0 9.7
3.9 –24.1
⬍.0001
2.0 –12.2
⬍.001
1.0a 4.9
Abbreviation: CI, confidence interval. a Because the risk of persistent disease at last follow-up in the final model was not statistically different between pNx, pN0 and pN1 microscopic, all these groups were pooled and set as the reference category.
be excluded that nodal involvement could have been missed in a few patients falsely considered as cN0 patients. A multivariate analysis confirmed that the presence of macroscopic nodal involvement was the most powerful predictor (RR 4.5) of PRD but that microscopic nodal involvement was also identified as an independent risk factor of PRD (RR 2.5), in contrast to our first study (2). This difference is probably related to the SPECT-CT’s capacity to detect small-volume lesions in or outside the neck, as previously discussed. The present data show, however, that only small nodal metastases (2–10 mm), but not micrometastases (⬍2 mm), were found to predict PRD. A limited number of patients in each subgroup may be an explanation for those statistical results, even if there is a rationale for a correlation between the nodal size and the risk of PRD including the pN1 microscopic patients. Moreover, the multivariate analyses using the N1a/N1b classification, or the nodal ECE, provide other evidence for this relationship. Assuming that the proportion of microscopic nodal involvement is equivalent in the pNx and pN1 microscopic patients, because both are cN0, we could have expected a
similar rate of PRD for these two categories of patients. However, Nx patients had a better outcome than the N1 microscopic patients. This suggests a selection bias for LND and is in accordance with the lower proportion of extrathyroid extension in pNx than in pN1 microscopic patients (6% vs 32%), the only prognosis-related difference. Indeed, a multivariate analysis confirmed that extrathyroid extension remained a strong (RR 2.6) risk factor for PRD as previously shown (2, 15). In addition, tumor size larger than 20 mm, but not male gender, was found as a significant risk factor (RR 3.4) in accordance with earlier reports (5, 19). In the present study, we did not include the LND modalities into the multivariate model because we believe that the benefits of prophylactic LND on locoregional PRD can be assessed only in prospective randomized trials. Data also demonstrate that the proportion of patients with NED at the last visit decreased progressively from pNx-pN0 (98%–93%) and pN1 microscopic (89%) to pN1 macroscopic patients (70%). However, although macroscopic nodal involvement (OR 4.9) and extrathyroid extension (OR 9.7) were independent predictors for
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
doi: 10.1210/jc.2014-1199
jcem.endojournals.org
Table 4. Final Models of Multivariate Analysis on Structural PRD Using Four Variables Related to N Status Adjusted for Thyroid Tumor Size and Extrathyroid Extension Model
Patients at Risk, n
RR
95% CI
Table 5. Final Models of Multivariate Analysis on Persistent Structural or Biochemical Disease at Last Follow-Up Visit Using Four Variables Related to N Status Adjusted for Extrathyroid Extension
P Value
Model
A, using the size of the largest metastatic nodal focus pNx 128 1.0a pN0 84 ⱕ2 mm 19 2–10 mm 21 2.6 1.1– 6.3 ⬍.04 ⬎10 mm 49 3.8 2.0 –7.4 ⬍.0001 B, using the TNM classification pNx 128 1.0 pN0 84 N1a 49 3.0 1.4 – 6.4 ⬍.01 N1b 44 4.5 2.2–9.2 ⬍.0001 C, using the number of metastatic lymph nodes pNx 128 1.0 pN0 84 1–5 60 3.1 1.5– 6.4 ⬍.01 ⬎5 32 4.6 2.2–9.7 ⬍.0001 D, using the LN ECE pNx 128 1.0 pN0 84 No ECE 72 3.5 1.8 – 6.9 ⬍.001 ECE 20 4.2 1.8 –9.8 ⬍.01
Patients at Risk, n
OR
95% CI
Abbreviations: CI, confidence interval; LN, lymph node.
a
a
persistent disease, microscopic nodal involvement was not. This suggests that PRD associated with micrometastases was more easily curable than that associated with macroscopic nodal involvement. Interestingly, although the proportions of RAI refractory lesions were similar in pN1 microscopic or macroscopic patients, small-volume lesions tended to be more frequent in pN1 microscopic patients (55%) than in pN1 macroscopic patients (26%). Furthermore, treatment for PRD in pN1 macroscopic patients was intensified compared with pN1 microscopic patients, ie, surgery alone (35% vs 18%), RAI treatment alone (17% vs 36%), surgery and RAI treatment (26% vs 18%), and no additional treatment (13% vs 18%). Although the clinical significance of biochemical disease is variable, with some patients evolving toward a clinical recurrence and others showing a stable or slowly decreasing Tg level over time (20), the present data suggest a higher rate of residual disease at 4 years in pN1 microscopic and, even more, in pN1 macroscopic patients than in pNx-pN0 patients. This retrospective study presents several limitations, particularly due to the diversity of surgeons and surgical procedures. However, most patients were operated on in
P Value
A, using the size of the largest metastatic nodal focus pNx 128 1.0a pN0 84 ⱕ2 mm 19 2–10 mm 21 5.6 1.4 –22.5 ⬍.02 ⬎10 mm 49 5.5 2.0 –14.8 ⬍.001 B, using the TNM classification pNx 128 1.0 pN0 84 N1a 49 N1b 44 3.7 1.5–9.4 ⬍.01 C, using the number of metastatic lymph nodes pNx 128 1.0 pN0 84 1–5 60 3.2 1.1–9.0 ⬍.04 ⬎5 32 4.4 1.4 –14.3 ⬍.02 D, using the LN ECE pNx 128 1.0 pN0 84 No ECE 72 ECE 20 5.3 1.6 –17.2 ⬍.01
Abbreviations: CI, confidence interval; LN, lymph node. Because the risk of PRD was not statistically different between pNx, pN0, and 2 mm or less in model A and between pNx and pN0 in models B, C, and D, these groups were pooled and set as the reference category.
139
Because the risk of persistent disease at last follow-up visit was not statistically different between pNx, pN0, and 2 mm or less in model A, between pNx, pN0, and pN1a in model B, between pNx and pN0 in model C, and between pNx, pN0, and no ECE in model D, these groups were pooled and set as the reference category.
specialized institutions and most by head and neck surgeons who have shared similar training. Furthermore, the concept of microscopic/macroscopic LN involvement has been developed for several years and is generally adopted by the referring surgeons. The elaboration in 2009 of regional recommendations for thyroid cancer management also contributed to homogenize the indications and procedures of LND as well as the drafting of the surgical report. Therefore, we believe that we can confidently rely on the peroperative data of patients managed inside or outside our institution. The diversity of LND modalities could also account for differences in the risk of PRD observed in pN0 or pN1 patients. However, even if a similar proportion of pN0 or pN1 microscopic patients underwent a prophylactic LND, the risk of PRD was significantly higher in pN1 microscopic than in pN0 patients. Furthermore, even if pN1 macroscopic patients were submitted in general to a more extended LND than the pN1 microscopic patients, their risk of PRD remained significantly higher in comparison with pN1 microscopic patients. This reinforces our conclusion that pN0, pN1 microscopic, or pN1 macroscopic PTC patients present different levels of risk of PRD.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
140
Bardet et al
Microscopic Lymph Node Involvement
At a time when therapeutic deescalation is considered in PTC patients, our data suggest that recognizing microscopic nodal involvement in a cN0 patient remains an important issue. Despite potential complications, prophylactic central dissection is the only way to identify microscopic nodal involvement. Furthermore, post-operative RAI still may be indicated in N1 microscopic patients, at least in those with small nodal metastases (2–10 mm), for two reasons. First, post-RAI scintigraphy with SPECT-CT is likely to be the most reliable imaging modality to detect small iodine-avid residual lesions. Second, RAI treatment is the most appropriate tool to destroy such lesions (21). If randomized prospective studies are necessary to address the issue of the benefit of RAI in patients with or without microscopic nodal involvement, the diagnostic and therapeutic impacts of RAI might introduce some bias between the ablated and the nonablated group. In conclusion, the present study demonstrates that the risk of PRD is likely higher than expected with the use of sensitive imaging tools and that patients with initial microscopic lymph node involvement present an intermediate outcome between that observed in N0-Nx patients and N1 macroscopic patients. Although long-term follow-up is needed, these retrospective data provide a new basis for discussing the management of PTC patients and could be taken into account for modifications to the current risk recurrence staging systems.
J Clin Endocrinol Metab, January 2015, 100(1):132–140
4.
5. 6.
7.
8.
9.
10.
11.
12.
13.
14.
Acknowledgments We greatly thank Professor Martin Schlumberger (Institut Gustave Roussy, Villejuif, France) for his helpful discussion. We are indebted to Helen Lapasset for her assistance in reviewing the manuscript. Address all correspondence and requests for reprints to: Stéphane Bardet, MD, Department of Nuclear Medicine and Thyroid Unit, Centre François Baclesse, 3 Avenue Général Harris, BP 5026, F-14076 Caen, Cedex 05, France. E-mail: [email protected]. Disclosure Summary: The authors have nothing to declare.
15.
16.
17.
18.
19.
References 1. Randolph GW, Duh QY, Heller KS, et al. The prognostic significance of nodal metastases from papillary thyroid carcinoma can be stratified based on the size and number of metastatic lymph nodes, as well as the presence of extranodal extension. Thyroid. 2012;22: 1144 –1152. 2. Bardet S, Malville E, Rame JP, et al. Macroscopic lymph-node involvement and neck dissection predict lymph-node recurrence in papillary thyroid carcinoma. Eur J Endocrinol. 2008;158:551–560. 3. Ito Y, Tomoda C, Uruno T, et al. Ultrasonographically and anatomopathologically detectable node metastases in the lateral compart-
20.
21.
ment as indicators of worse relapse-free survival in patients with papillary thyroid carcinoma. World J Surg. 2005;29:917–920. Wada N, Duh QY, Sugino K, et al. Lymph node metastasis from 259 papillary thyroid microcarcinomas: frequency, pattern of occurrence and recurrence, and optimal strategy for neck dissection. Ann Surg. 2003;237:399 – 407. Edge SB, Byrd DR, Compton CC, Fritz AG, Green FL, Trotti A. AJCC Cancer Staging Manual. New York: Springer; 2010. Cooper DS, Doherty GM, Haugen BR, et al. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid. 2009;19:1167–1214. Aide N, Heutte N, Rame JP, et al. Clinical relevance of single-photon emission computed tomography/computed tomography of the neck and thorax in postablation (131)I scintigraphy for thyroid cancer. J Clin Endocrinol Metab. 2009;94:2075–2084. Grewal RK, Tuttle RM, Fox J, et al. The effect of posttherapy 131I SPECT/CT on risk classification and management of patients with differentiated thyroid cancer. J Nucl Med. 2010;51:1361–1367. Schmidt D, Szikszai A, Linke R, Bautz W, Kuwert T. Impact of 131I SPECT/spiral CT on nodal staging of differentiated thyroid carcinoma at the first radioablation. J Nucl Med. 2009;50:18 –23. Leboulleux S, El Bez I, Borget I, et al. Post-radioiodine treatment whole body scan in the era of fluorodesoxyglucose positron emission tomography for differentiated thyroid carcinoma with elevated serum thyroglobulin levels. Thyroid. 2012;22:832– 838. De Lellis RA, Lloyd RV, Heitz PU, Eng C, eds. Word Health Organization Classification of Tumours. Pathology and Genetics of Tumours of Endocrine Organs. Lyon, France: IARC Press; 2004. Schlumberger M, Catargi B, Borget I, et al. Strategies of radioiodine ablation in patients with low-risk thyroid cancer. N Engl J Med. 2012;366:1663–1673. Tuttle RM, Tala H, Shah J, et al. Estimating risk of recurrence in differentiated thyroid cancer after total thyroidectomy and radioactive iodine remnant ablation: using response to therapy variables to modify the initial risk estimates predicted by the new American Thyroid Association staging system. Thyroid. 2010;20:1341–1349. Schlumberger M, Sherman SI. Approach to the patient with advanced differentiated thyroid cancer. Eur J Endocrinol. 2012;166:5–11. Leboulleux S, Rubino C, Baudin E, et al. Prognostic factors for persistent or recurrent disease of papillary thyroid carcinoma with neck lymph node metastases and/or tumor extension beyond the thyroid capsule at initial diagnosis. J Clin Endocrinol Metab. 2005; 90:5723–5729. Machens A, Dralle H. Correlation between the number of lymph node metastases and lung metastasis in papillary thyroid cancer. J Clin Endocrinol Metab. 2012;97:4375– 4382. Vas Nunes JH, Clark JR, Gao K, et al. Prognostic implications of lymph node yield and lymph node ratio in papillary thyroid carcinoma. Thyroid. 2013;23:811– 816. Verburg FA, Mader U, Tanase K, et al. Life expectancy is reduced in differentiated thyroid cancer patients ⱖ 45 years old with extensive local tumor invasion, lateral lymph node, or distant metastases at diagnosis and normal in all other DTC patients. J Clin Endocrinol Metab. 2013;98:172–180. Hay ID, Bergstralh EJ, Goellner JR, Ebersold JR, Grant CS. Predicting outcome in papillary thyroid carcinoma: development of a reliable prognostic scoring system in a cohort of 1779 patients surgically treated at one institution during 1940 through 1989. Surgery. 1993;114:1050 –1057. Baudin E, Do CC, Cailleux AF, Leboulleux S, Travagli JP, Schlumberger M. Positive predictive value of serum thyroglobulin levels, measured during the first year of follow-up after thyroid hormone withdrawal, in thyroid cancer patients. J Clin Endocrinol Metab. 2003;88:1107–1111. Schmidt D, Linke R, Uder M, Kuwert T. Five months’ follow-up of patients with and without iodine-positive lymph node metastases of thyroid carcinoma as disclosed by (131)I-SPECT/CT at the first radioablation. Eur J Nucl Med Mol Imaging. 2010;37:699 –705.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 02 February 2015. at 14:14 For personal use only. No other uses without permission. . All rights reserved.
