Late Relapses in Stage I Testicular Cancer Patients on Surveillance

Late Relapses in Stage I Testicular Cancer Patients on Surveillance

EURURO-6725; No. of Pages 7 EUROPEAN UROLOGY XXX (2016) XXX–XXX available at www.sciencedirect.com journal homepage: www.europeanurology.com Testis ...

340KB Sizes 0 Downloads 9 Views

EURURO-6725; No. of Pages 7 EUROPEAN UROLOGY XXX (2016) XXX–XXX

available at www.sciencedirect.com journal homepage: www.europeanurology.com

Testis Cancer

Late Relapses in Stage I Testicular Cancer Patients on Surveillance Mette Saksø Mortensen a,*, Jakob Lauritsen a, Maria Gry Gundgaard Kier a,b, Mikkel Bandak a, Ane Lindegaard Appelt a, Mads Agerbæk c, Niels Vilstrup Holm d, Mette Moe Kempel e, Hans von der Maase a, Gedske Daugaard a a

Danish Cancer Society, Copenhagen, Denmark;

c

Department of

Department of Oncology, Odense University Hospital, Odense, Denmark;

e

Department of

Department of Oncology, Copenhagen University Hospital, Copenhagen, Denmark;

Oncology, Aarhus University Hospital, Aarhus, Denmark;

d

b

Oncology, Aalborg University Hospital, Aalborg, Denmark

Article info

Abstract

Article history: Accepted March 6, 2016

Background: Comprehensive data on late relapse (LR) and very LR (VLR) in patients with clinical stage I (CS-1) testicular cancer followed on surveillance are missing. These data are essential for planning optimal follow-up. Objective: Assess incidence and outcome of LR (>2 yr) and VLR (>5 yr) in a large cohort of CS-1 surveillance patients, and examine differences in the clinical characteristics of patients with early relapse (ER), LR, and VLR. Design, setting, and participants: CS-1 surveillance patients diagnosed between 1984 and 2007 were identified from the retrospective Danish Testicular Cancer (DaTeCa) database. Outcome measurements and statistical analysis: We estimated survival and relapse probabilities and compared the results using log-rank tests and Cox regression analyses. We compared differences in patient characteristics by using x2, Fisher exact, and MannWhitney tests. Results and limitations: Our study included 3366 (2000 seminoma and 1366 nonseminoma) patients. Median follow-up was 15 yr. Five-year conditional risk of LR was 5.0% and 2.1% for seminoma and nonseminoma patients, respectively. There were no significant differences in disease-specific or overall survival when comparing the LR(VLR) and ER patients by log-rank, but Cox regression adjusted for age showed a significant effect of time to relapse on survival for seminoma patients. Apart from significantly more ER nonseminoma patients with elevated human chorionic gonadotropin at relapse, there were no significant differences in patient characteristics at orchiectomy or relapse. Limitations include retrospective design and exclusion of patients who had been offered adjuvant therapy. Conclusions: The risk of VLR is minimal, and the patients carry a good prognosis. Patient characteristics of CS-1 surveillance patients with LR(VLR) do not differ significantly from patients with ER. Patient summary: We compared stage I testicular cancer surveillance patients with early relapse (ER) versus late relapse (LR; >2 yr). LR patients as a group did no worse than ER patients, although increased time to relapse was negatively associated with survival for seminoma patients. # 2016 European Association of Urology. Published by Elsevier B.V. All rights reserved.

Associate Editor: Stephen Boorjian Keywords: Late relapse Stage I germ cell cancer Surveillance Seminoma Nonseminoma

* Corresponding author. Department of Oncology 5073, Rigshospitalet, Blegdamsvej 9, 2100 Copenhagen, Denmark. Tel. +45 51701413; Fax: +45 35453990. E-mail address: [email protected] (M.S. Mortensen). http://dx.doi.org/10.1016/j.eururo.2016.03.016 0302-2838/# 2016 European Association of Urology. Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Mortensen MS, et al. Late Relapses in Stage I Testicular Cancer Patients on Surveillance. Eur Urol (2016), http://dx.doi.org/10.1016/j.eururo.2016.03.016

EURURO-6725; No. of Pages 7 2

EUROPEAN UROLOGY XXX (2016) XXX–XXX

1.

programme. Disease-specific survival (DSS) and overall survival (OS)

Introduction

after relapse were calculated from date of first relapse until death from

Late relapse (LR) >2 yr after primary successful treatment for testicular cancer (TCa) is a rare event. Data on clinical stage I (CS-1) TCa patients on surveillance are sparse. The rate of LR for these patients is estimated to be 1–3% [1–4]. Most studies on LR for CS-1 patients are hampered by short follow-up, nonconsecutive series of patients, and lack of descriptions of clinical outcome. Accordingly, the primary aim of the present study was to assess the incidence and clinical outcome for LR chemotherapy-naı¨ve patients in a large unselected cohort of CS-1 patients undergoing surveillance. The secondary aim was to examine differences in risk factors in patients with early relapse (ER), LR, and very LR (VLR).

TCa or treatment complications (DSS) or until death from any cause (OS). We calculated median follow-up using the reverse–Kaplan-Meier method, with the status indicator (relapse) reversed [11]. To examine the specific relationship between time to relapse and DSS and OS, we constructed Cox proportional hazard regression models. The models included time from orchiectomy to relapse and age at relapse, with both variables added to the model using restricted cubic splines, with three to four knots, to account for nonlinearity. Because of the small number of VLR cases, we merged the LR and VLR groups into one that held all patients relapsing after 24 mo for subsequent analyses: LR(VLR). Survival after relapse was compared between patients with ER and patients with LR(VLR) using log-rank statistics. To assess our second aim, differences in patient characteristics for ER and LR(VLR), we used x2 test or Fisher exact test for categorical variables and Mann-Whitney U test for continuous variables. We compared the

2.

Patients and methods

following characteristics: age, tumour size, vascular invasion, invasion of rete testis, invasion of epididymis, AFP and hCG elevation at time of

From the Danish Testicular Cancer (DaTeCa) database (Supplement 1), identified

we

all

CS-1

patients

undergoing

surveillance

from

1984 through 2007 (n = 3774). Patients receiving adjuvant treatment

orchiectomy and at time of relapse as well as presence of embryonic carcinoma, endodermal sinus tumour, chorionic carcinoma, and teratoma at time of diagnosis for nonseminoma patients. All analyses were carried out separately for seminoma and

(n = 373) and patients who had synchronous bilateral TCa (n = 35) were excluded, leaving 3366 patients for the present analysis.

nonseminoma patients. The p values were two-sided and considered

All patients underwent primary inguinal orchiectomy followed by

significant when p < 0.05. Statistical analyses were performed using

staging, with measurements of tumour markers (TMs): a-fetoprotein

SPSS version 22.0 (IBM Corp., Armonk, NY, USA) and R version 2.15.3

(AFP), human chorionic gonadotropin (hCG), lactate dehydrogenase

(R Foundation for Statistical Computing, Vienna, Austria).

(LDH), computed tomography (CT) scan of the abdomen, and thoracic xray or CT scan. Patients were offered 5 yr of follow-up after orchiectomy and were included in the analyses irrespective of their adherence to the surveillance programme. A core biopsy was performed at relapse except for some cases with clear progression on CT scan or significant increase in TMs. Patients with

3.

Results

The final cohort consisted of 2000 patients with seminoma and 1366 patients with nonseminoma TCa.

relapse were offered radiation therapy (seminoma stage IIA/B) or chemotherapy (bleomycin, etoposide, and cisplatin [BEP]). In the case of

3.1.

Seminoma patients

3.1.1.

Incidence and outcome

residual tumour after chemotherapy, surgical removal of the tumour was performed. Clinical data from the time of orchiectomy and relapse were obtained from patient files and included in the DaTeCa database. Information about the DaTeCa database, including data on prognostic factors for relapse and detection of relapses, are thoroughly described in two previous papers [5,6]. Through links with several Danish registries [7–10], we identified patients with relapse after termination of the follow-up programme and patients lost to follow-up. Data were updated 30 November 2012. This study uses the following definitions of time to relapse: (1) ER refers to relapse 0–24 mo after orchiectomy; (2) LR refers to relapse 25–60 mo after orchiectomy; and (3) VLR refers to relapse after the patient had completed the 60-mo follow-up programme.

2.1.

Statistics

Our primary aim was to access the incidence and clinical outcomes of patients with LR. To estimate these outcomes, we calculated relapse and survival probabilities by using the Kaplan-Meier method. Patients were censored at the time of emigration (n = 67) if lost from the Danish civil registers (n = 4), at time of a metachronous TCa (n = 70), or at time of linkage to the national registries (30 November 2012). Time to relapse was defined as the time from the date of orchiectomy to the date of relapse diagnosis. We used conditional survival estimates to assess the 5-yr risk of LR, calculating the 5-yr cumulative incidence of relapse given that the patients were relapse free for the first 2 yr of follow-up. Conditional 10-yr risk of VLR was calculated as cumulative incidence given that the patients were relapse free during the 5-yr follow-up

Median follow-up for patients was 15 yr (interquartile range [IQR]: 9–21). A total of 388 patients relapsed, with a median time to relapse of 13 mo (IQR: 7–25); 288 had ER, 83 had LR, and 17 relapsed after termination of the 5-yr surveillance programme (VLR). The 2-yr cumulative risk of ER was 14.5% (95% confidence interval [CI],13.0–16.9) (Fig. 1). The 5-yr conditional risk of LR was 5.0% (95% CI, 4.0–6.1), while the 10-yr conditional risk of VLR was 1.0% (95% CI, 0.6–1.7). Twelve seminoma patients died of TCa or treatment complications: Four ER patients, one LR patient, and two VLR patients died of disseminated disease, while three ER patients, one LR patient, and one VLR patient died of treatment complications. One of the two VLR patients had nonseminoma histology in the relapse (Table 1). Comparing ER with LR(VLR) patients by log-rank test, the two groups demonstrated no significant differences in DSS (p = 0.13) or OS (p = 0.5). For patients with LR(VLR), 10-yr DSS and OS were 94.9% (95% CI, 88.1–97.8) and 89.3% (95% CI, 80.2–94.4), respectively. Increasing time to relapse was significantly associated with worse DSS and OS after relapse in an approximately linear fashion (p values for linear effects were 0.03 and 0.01, respectively; p values for nonlinear effects were 0.08 and 0.18). In the Cox regression model, 10-yr DSS after relapse at

Please cite this article in press as: Mortensen MS, et al. Late Relapses in Stage I Testicular Cancer Patients on Surveillance. Eur Urol (2016), http://dx.doi.org/10.1016/j.eururo.2016.03.016

EURURO-6725; No. of Pages 7 3

EUROPEAN UROLOGY XXX (2016) XXX–XXX

Primary histology

40

Cumulative incidence (%)

Nonseminoma Seminoma

30

3.1.2. 20

10

0 0

5

10

15

20

25

Time to relapse (years)

No. at risk Nonseminoma 1366 Seminoma

respectively. The small number of events following VLR did result in very uncertain estimates of the effect of late time to relapse on survival, however.

2000

913

692

492

292

130

1574

1151

819

464

155

Fig. 1 – Cumulative incidence of relapse from time of orchiectomy.

2 and 5 yr were 98.5% (95% CI, 96.7–100) and 96.3% (95% CI, 92.8–99.9), respectively, for a patient 38 yr of age at the time of relapse. The 10-yr OS after relapse at 2 and 5 yr were 98.2% (95% CI, 96.9–99.5) and 96.9% (95% CI, 94.6–99.3),

Differences in patient characteristics

Table 2 shows patient characteristics at the time of orchiectomy, and Table 3 summarises the characteristics and treatment at time of relapse. Univariate analyses of patient characteristics at the time of orchiectomy and at the time of relapse showed no significant differences between the ER and LR(VLR) groups (Supplementary Table 1). Eight of the LR incidents were detected on the final CT scan at 5-yr follow-up. All VLR cases were detected through patient symptoms (stomach and back pain and enlarged cervical lymph nodes). Details of the 17 VLR cases are shown in Table 1. The hCG level was increased in eight (47%) of these patients. Biopsies of their relapses revealed seminoma, and none of the eight had elevated TMs at the time of orchiectomy. All relapsing patients were in the good prognosis group, apart from two patients with LR, who were in the intermediate prognosis group (one because of an hCG level of 6975 IU/l and nonseminoma histology, the other because of bone metastases).

Table 1 – Very late relapses: patient characteristics at orchiectomy (primary) and at the time of relapse ID

Time to relapse, yr

Age at relapse, yr

Contralateral biopsy¤

250 369 476 1076 2352 2687 2709 2719 3113

9.5 14.3 9.7 9.3 6.3 11.4 7.4 7.2 6.4

56 45 51 37 49 33 43 31 84

No CIS No CIS No CIS No CIS No CIS NA No CIS CIS NA

3728 4155 4163

9.8 7.5 6.6

54 39 55

4172 4576 4865 5256 5384 133 633

10.4 5.4 10.9 7.0 6.4 8.8 14.5

4091

4331

AFP, primary/ relapse

hCG, primary/ relapse

Histology, primary/ relapse

Relapse stage

–/– –/– –/– –/– –/– –/– –/– –/– –/–

–/24 –/– –/22 –/72 –/56 13/– 260/– –/19.8 –/–

IIc IIc IIc IIc IIc IIc IIc IIc IIb

No CIS No CIS No CIS

–/– –/– –/–

–/116 –/36 –/–

S/S S/S S/S S/no biopsy S/S S/NS + T S/NA** S/S S/no biopsy S/S S/S S/S

40 57 42 38 29 43 67

No CIS NA No CIS NA NA No CIS No CIS

–/– –/27 –/– –/– –/– –/– –/–

–/26 921 –/– 60/– –/– –/– –/– –/–

8.8

43

NA

–/370

–/–

25.7

52

No CIS

4100/–

230/–

Treatment

Posttreatment surgery

Second relapse

Status

BEP BEP BEP BEP BEP BEP BEP BEP RT

Necrosis Necrosis Necrosis Necrosis – EC + T Necrosis – –

– – – – – Yes – – –

NED NED NED NED NED DoD NED NED NED

IIc IIIa IIa

BEP BEP Surgery*

– – –

– – Yes

NED NED DoT

S/S S/NS S/S S/S S/S EC/EC S+T/S

IIIa IIc IIc IIIa IIc IIIb IIc

BEP BEP BEP BEP BEP BEP BEP

– – – – – – Necrosis#

– – – – Yes – –

DoC DoC NED NED DoD NED DoC

S + EC + EST/no biopsy S + EC + EST/NS

IIb

BEP

T



NED

IIc

BEP

Fibrosis



NED

Cause of death

TCa

Pulmonary oedema Angiosarcoma Lung cancer

TCa Prostate cancer

AFP = a-fetoprotein; BEP = bleomycin, etoposide, and cisplatin; CIS = carcinoma in situ; DoC = deceased due to other cancer; DoD = deceased due to disease; DoT = deceased due to treatment complications; EC = embryonal carcinoma; EST = endodermal sinus tumour; hCG = human chorionic gonadotropin; NA = not available; NED = no evidence of disease; NS = nonseminoma; RT = radiation therapy; S = seminoma; T = teratoma; TCa = testicular cancer. ¤ Part of the Danish CIS-screening programme. CIS in contralateral testicle at the time of orchiectomy; the patient received RT (16 Gy) for his CIS. * Surgery for local relapse in scrotum; the patient received BEP for his second relapse but died of pulmonary oedema during BEP-treatment. ** Inconclusive biopsy. # Biopsies of residual tumour only.

Please cite this article in press as: Mortensen MS, et al. Late Relapses in Stage I Testicular Cancer Patients on Surveillance. Eur Urol (2016), http://dx.doi.org/10.1016/j.eururo.2016.03.016

EURURO-6725; No. of Pages 7 4

EUROPEAN UROLOGY XXX (2016) XXX–XXX

Table 2 – Patient characteristics at the time of orchiectomy Seminoma Characteristics Age, yr, median (IQR) Tumour size, median, cm* (IQR) Vascular invasion, n (%) Yes No Missing Invasion of rete testis, n (%) Yes No Missing Invasion of epididymis, n (%) Yes No Missing hCG elevation, n (%) Yes No Missing AFP elevation, n (%) Yes No Missing Embryonal carcinoma, n (%) Yes No Endodermal sinus tumour, n (%) Yes No Chorionic carcinoma, n (%) Yes No Teratoma, n (%) Yes No

Nonseminoma

ER 37 (31–44) 4.0 (3.0–5.5)

LR(VLR) 35 (31–45) 3.5 (2.0–5.5)

ER 29 (25–36) 3.0 (2.0–5.0)

LR (VLR) 32 (28–38) 3.0 (2.5–5.0)

54 (19) 127 (44) 107 (37)

8 (10) 37 (44) 38 (42)

102 (25) 156 (39) 142 (36)

5 (25) 10 (50) 5 (25)

121 (42) 99 (34) 68 (24)

36 (43) 32 (39) 15 (18)

130 (32) 148 (37) 122 (30)

4 (20) 11 (55) 5 (25)

36 (13) 204 (71) 48 (17)

10 (12) 63 (76) 10 (12)

23 (6) 271 (68) 106 (26)

1 (5) 17 (85) 2 (10)

35 (12) 240 (83) 13 (5)

9 (11) 69 (83) 5 (6)

159 (40) 220 (55) 21 (5)

4 (20) 16 (80) 0 (0)

0 (0) 278 (97) 10 (3)

0 (0) 78 (94) 5 (6)

199 (50) 183 (46) 18 (4)

7 (35) 13 (65) 0 (0)





365 (91) 35 (9)

20 (83) 4 (17)





146 (37) 254 (63)

6 (25) 18 (75)





70 (17) 330 (83)

2 (8) 22 (92)





206 (51) 194 (49)

8 (33) 16 (67)

AFP = a-fetoprotein; ER = early relapse; hCG = human chorionic gonadotropin; IQR = interquartile range; LR(VLR) = late relapse and very late relapse. Exact tumour size was available for 1688 seminoma patients and 739 nonseminoma patients.

*

3.2. 3.2.1.

Nonseminoma patients Incidence and outcome

Median follow-up was 15 yr (IQR: 10–22). In total, 424 patients relapsed; 400 had ER, 20 had LR, and 4 had VLR. Median time to relapse was 5 mo (IQR: 3–10). The 2-yr cumulative risk of ER was 29.4% (95% CI, 27.1– 31.9) (Fig. 1). The 5-yr conditional cumulative risk of LR was 2.1% (95% CI, 1.4–3.3), and the 10 yr cumulative risk of VLR was 0.3% (95% CI, 0.1–1.1). In total, 15 of the nonseminoma patients died of TCa or treatment complications: Nine ER patients died of progressive disease, while five ER patients and one LR patient died of treatment complications. There was no significant effect of time to relapse on DSS or OS after relapse (all p > 0.1), and visual inspection of the estimated relationships between time to relapse and DSS or OS showed no apparent effect, either. Similarly, log-rank comparison of ER versus LR(VLR) was nonsignificant for both DSS (p = 0.8) and OS (p = 0.8). For LR(VLR), 10-yr DSS and OS were 95.8% (95% CI, 73.9–99.4) and 89.8% (95% CI, 64.3–97.4), respectively.

3.2.2.

Differences in patient characteristics

Table 2 shows patient characteristics at the time of orchiectomy, and Table 3 summarises the characteristics at the time of relapse. There were no significant differences in primary histology between the two groups. Likewise, we detected no significant differences in other patient characteristics between the ER and LR(VLR) groups at the time of diagnosis. At the time of relapse, more patients in the ER group than in the LR(VLR) group had increased levels of hCG (p = 0.001) (Supplementary Table 1). Transformed teratoma was seen in two ER patients’ residual tumours after BEP treatment. Five of the LR cases were detected on the final 5-yr follow-up CT scan. All VLR cases were detected because of patients’ symptoms (stomach pain, back pain, and shortness of breath). Details of the four patients with VLR are shown in Table 1. Although the vast majority of relapses were in the good prognosis group, 3 (15%) of the LR patients and 17 (4%) of the ER patients were in the intermediate prognosis group. Poor prognosis was seen in only three ER patients. Details of treatment are found in Table 3.

Please cite this article in press as: Mortensen MS, et al. Late Relapses in Stage I Testicular Cancer Patients on Surveillance. Eur Urol (2016), http://dx.doi.org/10.1016/j.eururo.2016.03.016

EURURO-6725; No. of Pages 7 5

EUROPEAN UROLOGY XXX (2016) XXX–XXX

Table 3 – Patient characteristics at the time of relapse (patients grouped according to their primary histology)

Relapse characteristics No. of patients Age at relapse, yr, median (IQR) Time of relapse Year 0 Year 1 Year 2 Year 3 Year 4 >5 yr hCG elevation, n (%) Yes No AFP elevation, n (%) Yes No Relapse biopsy, n (%) Seminoma Nonseminoma Teratoma NA Relapse stage*, n (%) IIa IIb IIc IIIa IIIb IS NA Treatment, n (%) RT BEP Surgery only Post-treatment surgery, n (%) Teratoma Death from TCa, n Progressive disease Complications** 5-yr DSS, % 5-yr OS, %

Seminoma

Nonseminoma

Relapse group

Relapse group

ER

LR(VLR)

ER

LR(VLR)

288 38 (32–45)

100 39 (34–49)

400 29 (25–37)

24 39 (31–47)

169 119 – – – –

– – 45 21 17 17

338 62 – – – –

33 (11) 255 (89)

16 (16) 84 (84)

170 (43) 230 (57)

2 (8) 22 (92)

4 284 191 177 6 1 8

(1) (99)

(4) (96)

(14)

6 18 15 5 10 1 0

(25) (75)

(4)

178 222 133 8 106 14 19

(44) (56)

(4)

4 96 87 78 5 1 4

102 118 47 6 7 7 1

(35) (41) (16) (2) (2) (2) (0)

24 35 32 6 3 0 0

(24) (35) (32) (6) (3) (0) (0)

117 92 24 10 69 68 20

(29) (23) (6) (3) (17) (17) (5)

3 7 7 0 6 1 0

(13) (29) (29) (0) (25) (4) (0)

198 88 2 11 1

(69) (31) (1) (4)

42 57 1 14 2

(42) (57) (1) (14)

4 386 8 91 59

(1) (97) (2) (23)

0 21 3 8 1

(0) (87) (13) (33)

(93) (3)

4 3 98.2 95.8

(90) (6)

3 2 94.9 93.8

– – 8 5 7 4

(6) (80)

9 5 97.0 95.2

(33) (67) (0)

0 1 98.2 95.8

AFP = a-fetoprotein; BEP = bleomycin, etoposide, and cisplatin; DSS = disease-specific survival; ER = early relapse; hCG = human chorionic gonadotropin; IQR = interquartile range; IS = in situ; LR(VLR) = late relapse and very late relapse; NA = not available; OS = overall survival; RT = radiation therapy; TCa = testicular cancer. * See Supplementary Table 2 for definition of stages. ** Death from complications were kidney failure, pulmonary embolism, pulmonary oedema, septicaemia (n = 4), pneumonia (n = 2),and bleomycin-induced pneumonitis (n = 2).

4.

Discussion

In this large study of 3366 CS-1 patients followed on surveillance, we found a conditional 5-yr cumulative risk of LR of 5.0% for seminoma patients and 2.1% for nonseminoma patients. Comparing the two groups by log-rank, there were no significant differences in DSS or OS for ER versus LR(VLR) patients. For seminoma patients, however, we found a significant association between time to relapse, considered as a continuous variable, and DSS and OS. We found no significant differences in patient characteristics at the time of orchiectomy and at the time of relapse in ER versus LR(VLR) apart from more ER nonseminoma patients with elevated hCG levels at the time of relapse.

With more than 15 yr of follow-up and 124 LRs, the present study adds substantial information about both patient characteristics and survival for LR and VLR surveillance patients. We have not been able to identify any longterm studies with comparable data on the incidence of LR. The largest study of LR in patients on surveillance included 28 patients and was not able to calculate the incidence of LR [12]. A study from Memorial Sloan Kettering Cancer Center reported a 5-yr cancer-specific survival of 93% in 18 chemotherapy-naı¨ve patients with LR [13]. This finding is in line with our results. Fedyanin et al analysed data on 169 relapsing chemotherapy-naı¨ve stage I patients, including 29 patients with LR [14]. They found significantly worse

Please cite this article in press as: Mortensen MS, et al. Late Relapses in Stage I Testicular Cancer Patients on Surveillance. Eur Urol (2016), http://dx.doi.org/10.1016/j.eururo.2016.03.016

EURURO-6725; No. of Pages 7 6

EUROPEAN UROLOGY XXX (2016) XXX–XXX

outcome for seminoma patients with LR: Three-year OS was 91% and 65% for ER and LR, respectively. There were no significant differences in OS for the nonseminoma patients. In the present study, three of the VLR seminoma patients died of progressive disease or complications. All were in the good prognosis group, and none of them had any comorbidity that could explain the outcome, although one of them had nonseminoma histology in the relapse. Increasing time to relapse resulted in reduced DSS and OS for seminoma patients, but with only 5 disease-related deaths and 13 deaths overall among the LR(VLR) seminoma patients, these results regarding effects of time to relapse should be interpreted with caution. A large part of the VLR seminoma patients (47%) had elevated hCG levels at the time of relapse. This finding is in line with a study of surgical management of LR in CS-1 patients on surveillance, which also identified hCG as the predominant TM at relapse [12], although only two of the patients in that study had seminoma. A study of LR in nonseminoma CS-I patients undergoing surveillance concluded that seminoma was the predominant histology at relapse in 56% of nine relapses [4]. We could not confirm this finding in the present study. Nonseminoma histology was predominant in most relapse biopsies (67%) of nonseminoma patients in the LR(VLR) group. Various studies advocate for long, even lifelong, followup for TCa surveillance patients [13,15–18]. In the present study, the 10-yr cumulative risk of relapse for seminoma patients without relapse during the 5-yr follow-up programme was 1%. Accordingly, in a surveillance setting with annual CT scans, approximately 100 patients would need an annual scan for 5 yr to detect one relapse. This is not only costly but also adds a significant burden to the patients and the hospitals. The potential risk of second cancers after repeating CT scans in this group of patients is controversial [19–21]. In addition, the risk of dying of other causes will far exceed the risk of dying of TCa for these individuals. Hence, we agree with the authors of the Kollmannsberger study [1] that extending surveillance imaging schedules for all for the benefit of the few may not add overall value. The few relapses after 5-yr follow-up were all diagnosed because of patient symptoms. Consequently, it is of utmost importance to educate patients and general practitioners on the risk of VLR to avoid treatment delay. Measurement of TMs beyond 5-yr follow-up in patients with a history of CS-1 might be valuable but is not sufficient on its own because half of the VLR patients did not have elevated TMs at the time of relapse in the present study. The retrospective nature of our study adds some limitations. Early in the study period, 384 seminoma patients (most with tumour size >6 cm) received adjuvant treatment (not included in the present study). Thus, the relapse rate for seminoma patients could possibly have been higher with these patients included. In contrast, this study is by far the largest of LR in CS-1 surveillance patients. The long follow-up and population-based nature of the study add strength to the applicability to other surveillance populations.

5.

Conclusions

Chemotherapy-naı¨ve surveillance patients with LR(VLR) have a good prognosis, and their characteristics do not differ from patients with ER. We believe that patients with LR(VLR) can be managed safely by following the general recommendations for relapsing CS-1 patients [22]. The conditional risk of VLR is minimal; nevertheless, patients and physicians should be aware of this risk to minimise treatment delay. Future studies should focus on the optimal follow-up programme for CS-1 patients. Author contributions: Mette Saksø Mortensen had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Mortensen, Lauritsen, Kier, Bandak, von der Maase, Daugaard. Acquisition of data: Mortensen, Lauritsen, Kier, Bandak, von der Maase, Daugaard, Agerbæk, Holm, Kempel. Analysis and interpretation of data: Mortensen, Lauritsen, Kier, Bandak, von der Maase, Daugaard, Appelt. Drafting of the manuscript: Mortensen, Lauritsen, Kier, Bandak, von der Maase, Daugaard. Critical revision of the manuscript for important intellectual content: Mortensen, Lauritsen, Kier, Bandak, von der Maase, Daugaard, Agerbæk, Holm, Kempel, Appelt. Statistical analysis: Mortensen, Lauritsen, Bandak, Appelt. Obtaining funding: Mortensen, von der Maase, Daugaard. Administrative, technical, or material support: None. Supervision: von der Maase, Daugaard. Other (specify): None. Financial disclosures: Mette Saksø Mortensen certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None. Funding/Support and role of the sponsor: This work was supported by grants from The Danish Cancer Society, The Danish Cancer Research Foundation, and the Preben & Anna Simonsens Foundation. The grants were used only to pay a salary to Mette Saksø Mortensen during her PhD study of stage I germ cell cancer. The organisations were not involved in the study design, management, or data interpretation.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. eururo.2016.03.016. References [1] Kollmannsberger C, Tandstad T, Bedard PL, et al. Patterns of relapse in patients with clinical stage I testicular cancer managed with active surveillance. J Clin Oncol 2015;33:51–7. [2] Chung P, Parker C, Panzarella T, et al. Surveillance in stage I testicular seminoma—risk of late relapse. Can J Urol 2002;9:1637–40. [3] Boyer MJ, Cox K, Tattersall MH, Findlay MP, Grygiel J, Rogers J. Active surveillance after orchiectomy for nonseminomatous testicular germ cell tumors: late relapse may occur. Urology 1997;50:588–92.

Please cite this article in press as: Mortensen MS, et al. Late Relapses in Stage I Testicular Cancer Patients on Surveillance. Eur Urol (2016), http://dx.doi.org/10.1016/j.eururo.2016.03.016

EURURO-6725; No. of Pages 7 EUROPEAN UROLOGY XXX (2016) XXX–XXX

7

[4] Nolan L, Wheater M, Kirby J, Simmonds P, Mead G. Late relapse

[13] Sharp DS, Carver BS, Eggener SE, et al. Clinical outcome and pre-

(>2 years) on surveillance in stage I non-seminomatous germ cell

dictors of survival in late relapse of germ cell tumor. J Clin Oncol

tumours; predominant seminoma only histology. BJU Int 2010;106: 1648–51. [5] Mortensen MS, Lauritsen J, Gundgaard MG, et al. A nationwide cohort study of stage I seminoma patients followed on a surveillance program. Eur Urol 2014;66:1172–8. [6] Daugaard G, Gundgaard MG, Mortensen MS, et al. Surveillance for stage I nonseminoma testicular cancer: outcomes and long-term follow-up in a population-based cohort. J Clin Oncol 2014;32: 3817–23. [7] Storm HH, Michelsen EV, Clemmensen IH, Pihl J. The Danish Cancer Registry—history, content, quality and use. Dan Med Bull 1997; 44:535–9. [8] Bjerregaard B, Larsen OB. The Danish Pathology Register. Scand J Public Heal 2011;39(Suppl 7):72–4. [9] Helweg-Larsen K. The Danish Register of Causes of Death. Scand J Public Heal 2011;39(Suppl 7):26–9. [10] Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scand J Public Heal 2011;39(Suppl 7):30–3. [11] Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time. Control Clin Trials 1996;17:343–6. [12] Rice KR, Beck SD, Pedrosa JA, Masterson TA, Einhorn LH, Foster RS. Surgical management of late relapse on surveillance in patients presenting with clinical stage I testicular cancer. Urology 2014;84: 886–90.

2008;26:5524–9. [14] Fedyanin M, Tryakin A, Kanagavel D, et al. Late relapses (>2 years) in patients with stage I testicular germ cell tumors: predictive factors and survival. Urol Oncol 2013;31:499–504. [15] George DW, Foster RS, Hromas RA, et al. Update on late relapse of germ cell tumor: a clinical and molecular analysis. J Clin Oncol 2003;21:113–22. [16] Oldenburg J, Wahlqvist R, Fossa SD. Late relapse of germ cell tumors. World J Urol 2009;27:493–500. [17] Baniel J, Foster RS, Einhorn LH, Donohue JP. Late relapse of clinical stage I testicular cancer. J Urol 1995;154:1370–2. [18] Sheinfeld J, Feldman DR. Editorial comment. Urology 2014;84: 890–1. [19] van Walraven C, Fergusson D, Earle C, et al. Association of diagnostic radiation exposure and second abdominal-pelvic malignancies after testicular cancer. J Clin Oncol 2011;29:2883–8. [20] Brenner DJ, Hall EJ. Computed tomography—an increasing source of radiation exposure. N Engl J Med 2007;357:2277–84. [21] Brenner DJ, Shuryak I. Ten years of follow-up is not long enough to assess lifetime cancer risks caused by computed tomography scans in a young population. J Clin Oncol 2011;29:4062, author reply 4062. [22] Albers P, Albrecht W, Algaba F, et al. Guidelines on testicular cancer: 2015 update. Eur Urol 2015;68:1054–68.

Please cite this article in press as: Mortensen MS, et al. Late Relapses in Stage I Testicular Cancer Patients on Surveillance. Eur Urol (2016), http://dx.doi.org/10.1016/j.eururo.2016.03.016