Long-term clinical outcomes of patients with negative pathology (pN0) at primary retroperitoneal lymph node dissection

Abstract

Background

Patients who undergo primary retroperitoneal lymph node dissection (pRPLND) for early-stage testicular cancer and have no cancer (pN0) found in the retroperitoneum are believed to have an excellent prognosis. However, some experience relapse, potentially due to limitations of current staging methods. We aim to describe long-term outcomes and relapse patterns among a contemporary cohort of patients found to be pN0 at pRPLND to identify opportunities for improved diagnostic approaches and optimal patient selection.

Methods

We reviewed our prospectively maintained database for patients who underwent pRPLND for non-seminomatous germ cell tumors at our tertiary cancer center during the period from January 1, 2000, through September 30, 2023 (n=628). We excluded 282 patients with node-positive pathology for a final analytic cohort of 346 patients. Our primary outcome was recurrence-free survival (RFS). Secondary outcomes included timing and location of recurrence.

Results

Of 346 included patients with pN0 pathology, 23 experienced relapse with a two-year RFS rate of 93% (95% confidence interval: 90, 96). Most recurrences (70%) occurred in the lungs and within 6 months of pRPLND. Serum tumor markers were positive in 43% of patients at the time of relapse. All patients who relapsed were treated with salvage chemotherapy; 6 patients required additional surgical procedures. There were no testis cancer–related deaths.

Conclusions

Two-year RFS for patients with pN0 pRPLND pathology is excellent. All recurrences were outside of the retroperitoneum, suggesting subclinical distant metastases at time of surgery and the benefits of a bilateral template dissection. Improved diagnostics may help better identify patients with disease within or outside of the retroperitoneum prior to pRPLND, helping guide treatment decisions.

Introduction

Primary retroperitoneal lymph node dissection (pRPLND) is indicated for patients with clinical stage I (CSI) non-seminomatous germ cell tumors (NSGCT) with risk factors for relapse or select patients with serum tumor marker (STM) negative, low-volume retroperitoneal-only metastases (CSIIA/B).^1–5 In these situations, pRPLND serves both a diagnostic and therapeutic role, staging the retroperitoneum and curing 80–90% of patients while minimizing the burden of chemotherapy and its long-term side effects.^6–8 Inherent to the effectiveness of pRPLND is the predictable pattern of metastatic spread in testicular GCTs, where the retroperitoneum is the first site of extratesticular disease for most patients.

Clinical staging prior to pRPLND has limitations as 20–60% of patients will have no cancer found in the retroperitoneum (pN0).^{9, 10} Patients with pN0 pathology at pRPLND are assumed to have an excellent prognosis but have undergone an invasive procedure with morbidity and limited therapeutic benefit. Conversely, as diagnostic staging and patient selection have improved so has the potential for cure with pRPLND alone in patients with disease.⁷ Still, in patients with (pN+) and without (pN0) disease at pRPLND, relapses occur. Potential causes of relapse following pRPLND are either unresected disease in the retroperitoneum or subclinical metastases outside the retroperitoneum.

A contemporary assessment of patients who experience relapse following a negative (pN0) pRPLND can identify a group of patients who are either understaged by current diagnostic approaches or who exhibit atypical cancer behavior, both of whom would be less likely to benefit from pRPLND. A better understanding of factors associated with atypical relapse can help with patient selection, identify scenarios where implementation of novel biomarkers may be beneficial, and inform more personalized approaches to post-pRPLND surveillance. We describe patterns of relapse, explore factors associated with relapse, and report long-term clinical outcomes among a cohort of patients who underwent pRPLND and had no cancer found in the retroperitoneum.

Materials and Methods

Data Source and Patients

All patients treated for testicular GCTs at Memorial Sloan Kettering Cancer Center (MSK) are included in a prospectively maintained clinical database that collects patient demographics; diagnosis, staging, and treatment information, and pathology, radiology, and laboratory results. Long-term oncologic and survival outcomes are also collected and supplemented by death registry data.

Following institutional review board approval (IRB #16–554), we queried our database for all pRPLND procedures performed at MSK from January 1, 2000, through September 30, 2023. Patients who underwent pRPLND for pure seminoma were not included. Additional exclusions were patients who had non-GCT pathology at orchiectomy, those who were Clinical Stage IS, and those with nodal metastatic disease (pN+), including teratoma and/or other non-teratomatous GCT, at the time of pRPLND.

Measures and Outcomes

Patient-level variables included age at pRPLND, clinical stage (CS), primary tumor laterality, tumor histology, presence of lymphovascular invasion (LVI), and STM levels (alpha-fetoprotein [AFP], human chorionic gonadotropin [HCG], and lactate dehydrogenase [LDH]) prior to orchiectomy and pRPLND. The timing and results of preoperative cross-sectional imaging and any history of prior ipsilateral groin surgery (hernia repair, orchidopexy, torsion, etc) were also assessed. Lymph node measurements were abstracted from the radiology report. The greatest axial dimension of the largest lymph node was used for staging, per American Joint Committee on Cancer and National Comprehensive Cancer Network (NCCN) guidelines. Patients with any retroperitoneal lymph node ≥ 1 cm in the largest axial dimension were considered cN1/CSIIA. Since 2000 the standard approach to pRPLND at MSK has been an open bilateral template operation using nerve-sparing techniques when indicated for family planning or sexual health preservation. For staging, we perform imaging in close proximity to pRPLND with contrast-enhanced computed tomography (CT) of the chest, abdomen, and pelvis as well as repeat STM testing. Ipsilateral pelvic lymph node dissection was performed in the setting of equivocal or clinically suspicious upper pelvic lymph nodes on preoperative imaging or when encountered intraoperatively.

Our primary outcome of interest was recurrence-free survival (RFS), defined as freedom from recurrence or death from any cause. RFS was calculated as a time-to-event variable. Patients who did not die or relapse were censored at their last follow-up. Secondary outcomes included relapse locations and method of detection (radiographic vs. STMs [elevated AFP or HCG]), and patient and disease factors associated with relapse.

Statistical Analysis

Patient and disease characteristics were reported using descriptive statistics. RFS was calculated using the Kaplan–Meier method among all patients and compared by clinical stage, prior ipsilateral groin surgery, presence of LVI, and embryonal carcinoma at index orchiectomy. Cox regression modeling was used to determine if the following a priori chosen clinical factors were associated with relapse following pN0 pRPLND: age at RPLND, clinical stage at diagnosis, presence of embryonal carcinoma at index orchiectomy, vascular invasion at index orchiectomy, and elevated STMs prior to orchiectomy or time from last CT scan to RPLND. These factors were chosen for analysis based on previously reported associations with relapse following pRPLND. Multivariable analysis was not conducted due to low event frequency and instability of the adjusted model. All analyses were performed using Stata, version 15.0 (StataCorp, College Station, TX).

Results

pRPLND outcomes

During the study period 628 patients with NSGCT underwent pRPLND, and the 346 patients (55%) who did not have disease found in their retroperitoneum (pN0) comprised our final analytic cohort. Four hundred forty-six patients underwent pRPLND for CSI NSGCT and no cancer (pN0) was found in 295 (66%); the corresponding rate for the 182 patients with CSII disease found to have no cancer (pN0) was 28% (n=51). For patients with CSIIA disease, median preoperative clinical node size was 1.3 cm (interquartile range [IQR]: 1.1, 1.5). All patients with CSIIB (pre-RPLND lymph nodes >2 cm) had node-positive disease at RPLND (n=41). Of all 346 patients with pN0 pathology, 295 (85%) were CSI and 51 (15%) were CSII. Median time from orchiectomy to pRPLND for all patients was 45 days (IQR: 35, 63). Sixteen patients also had a concomitant ipsilateral pelvic lymph node dissection.

Most patients (85%) underwent nerve-sparing pRPLND performed by a single surgeon (JS, 92%). The median number of lymph nodes removed was 50 (IQR: 36, 64). Forty-four patients (13%) experienced a postoperative complication within 30 days of surgery, 80% of which were grade 1–2; the remainder were grade 3, most of which were related to chylous ascites or lymphocele.

Relapse Risk and Location

Twenty-three patients (21 CSI and 2 CSIIA) experienced relapse after pN0 pRPLND (Table 3), with a raw median time to relapse of 4.6 months (IQR: 3.0, 6.9; range: 2.3–84.6 months). One patient experienced a “late relapse” >2 years after pRPLND. Median follow-up time for patients who did not relapse or die was 6.0 years (IQR: 2.8, 11.0). For all patients who were pN0 at pRPLND, 2-year RFS was 93% (95% confidence interval [CI]: 90%, 96%). When stratified by clinical stage, 2-year RFS was 93% (95% CI: 89%,95%) for CSI and 96% (95% CI: 83%, 99%) for CSIIA. Of 19 patients (5.5%) who had a history of prior ipsilateral groin surgery, none experienced relapse (Supplemental Table 1). The most proximal cross-sectional imaging study prior to pRPLND was performed at a median of 15 days (IQR: 9, 23) prior to surgery for patients who did not relapse after pRPLND and 16 days (IQR: 10, 25) for patients who did experience relapse after pRPLND.

Table 3:

Relapse details of the 23 patients who were pN0 at primary RPLND and experienced relapse

	N=23
Raw Median Time to Relapse, months (IQR)	4.6 (3.0, 6.9)
Initial Clinical Stage
CSI	21 (91.3%)
CSII	2 (8.7%
Relapse Location
Lung only	9 (39%)
Markers only	2 (8.7%)
Pelvis	1 (4.3%)
Suprahilar	1 (4.3%)
Multiplea	10 (43%)
Relapse Histologyb
Embryonal carcinoma	3
Yolk sac tumor	2

Abbreviations: IQR, interquartile range; RPLND, retroperitoneal lymph node dissection.

^aMarkers + Pelvis (N=2); Markers + Lung (N=5); Markers + Suprahilar (N=1); Lung + Supraclavicular (N=1); Liver + Lung (N=1).

^bRelapse pathology prior to chemotherapy available for 5 patients.

On univariate Cox regression, age at pRPLND (HR 1.19 per 5 years, 95% CI 0.94, 1.49) and the presence of embryonal carcinoma in the orchiectomy specimen (HR 1.34, 95% CI 0.31, 5.70) had a higher hazard of recurrence. Patients with LVI (n=241, 70%) were more likely to experience relapse (HR 4.22, 95% CI: 0.99, 18.1; p=0.052), although the hazard ratio estimates for all factors lacked precision and statistical significance due to low recurrence rates overall, precluding further assessment with multivariable adjustment.

Anatomically, 16 of 23 patients (70%) experienced relapses in the lung; no relapses occurred in the retroperitoneum only (Table 3). STMs (AFP or HCG) were elevated at the time of relapse in 10 of 23 patients (43%), including two with relapsed detected solely by elevated STM levels. Seven patients underwent confirmatory biopsy of radiographically visible recurrent masses prior to additional treatment (embryonal carcinoma (3), yolk sac tumor (2), and unspecified germ cell tumor (2)).

Relapse treatment and outcomes

All patients with pN0 pathology were managed with surveillance after pRPLND with one exception: a patient who had secondary somatic malignancy in the orchiectomy specimen and underwent adjuvant chemotherapy due to concern about the high risk of extra-retroperitoneal recurrence.

All patients who experienced relapse were categorized as having good-risk disease (according to International Germ Cell Cancer Collaborative Group classification) and received first-line chemotherapy, most commonly (80%) four cycles of etoposide and cisplatin (EPx4). One patient experienced a second relapse after first-line treatment and underwent salvage high-dose chemotherapy with complete response. The remainder of the patients had a durable complete response following first-line chemotherapy.

Six patients underwent subsequent surgical resections (4 lung, 1 pelvic, and 1 mediastinal) after first-line chemotherapy. Two patients who underwent post-chemotherapy resection of lung masses had viable GCT, one with seminoma subsequently treated with surveillance and one with embryonal carcinoma who received 2 cycles of adjuvant EP. Both remain disease free. One patient who underwent post-chemotherapy resection of a pelvic mass had teratoma. Among all patients with pN0 pRPLND histology, there was one death, which was non–cancer related, in a patient who had not experienced relapse at last follow-up.

Discussion

In the modern era, the indications for pRPLND continue to expand with the aim of reducing long-term toxicities from chemotherapy or radiation among patients with decades of post-treatment life expectancy.^11–13 In this study we report long-term outcomes for patients undergoing pRPLND for NSGCT found to have no disease in the retroperitoneum (pN0). To our knowledge, this is the largest and most comprehensive series of pN0 patients reported and the only study to focus on timing, frequency, and patterns of recurrence in this unique population. Our results demonstrate a persistent need for improved diagnostics to optimize patient selection for pRPLND and the benefit of a complete bilateral template dissection, as recurrence in the retroperitoneum was common in a prior randomized study using modified template approaches.¹⁴

Patients undergoing pRPLND for CSI and CSII disease were pN0 in 66% and 28% of cases, respectively, reflecting the limitations of contemporary diagnostics (CT and traditional STM testing). These patients had excellent long-term outcomes with a 93% 2-year RFS and no testis cancer–related deaths. Identifying those patients at risk for distant recurrence after pN0 pRPLND would allow for better patient selection prior to surgery. However, due to low recurrence rates in our cohort, we were unable to identify any factors associated with an independent risk of relapse after pN0 pRPLND.

In our cohort, 66% of patients with CSI disease who underwent pRPLND had no disease in the retroperitoneum, which aligns with previously demonstrated rates of relapse (40–50%) in those managed with surveillance who have established risk factors.^{15, 16} We found that 28% of patients with CSIIA disease had no disease on pRPLND despite most patients having >1 cm retroperitoneal lymphadenopathy on staging imaging. This pN0 rate is higher than some published series and probably reflects the application of strict criteria for cN1 staging (largest axial dimension ≥1 cm) and potentially a Will Rogers phenomenon as patients with pN+ disease who had repeat imaging prior to surgery were more likely to be reclassified. While it is our practice to perform short-interval repeat imaging for patients with equivocal lymph nodes and normal STMs, there is concern that this may expose patients to risk of further progression and lower the likelihood of cure with pRPLND.^{17, 18} In a prior series from our institution, short interval surveillance for equivocal retroperitoneal lymph nodes prior to RPLND was associated with upstaging from CSI to CSII in 12% of patients.⁸

MicroRNAs (miRNAs) have been studied as a possible supplement to existing diagnostic modalities for patients with testicular GCT.^{19, 20} Patients with normalized miRNA-371 following radical orchiectomy may be at lower risk of relapse on surveillance than those with persistently elevated miRNA levels.¹⁹ Further, miRNA may serve to adjudicate small retroperitoneal masses and predict viable disease in the retroperitoneum prior to surgery.^{20, 21} Improved detection of residual disease may lower rates of pN0 pRPLND while also identifying those patients at risk for distant recurrence following pRPLND. Given that RPLND comes with some risk of long-term sexual dysfunction from retrograde ejaculation as well as incisional hernia and bowel obstruction, all efforts should be made to select patients for surgery with the highest risk of harboring retroperitoneal disease. A limitation of our current study is the lack of systematically collected patient-reported sexual health outcomes, including antegrade ejaculation rates, which is a critical component of survivorship.

Hematogenous or aberrant lymphatic drainage may result in “skip metastases” absent retroperitoneal disease (pN0) leading to atypical relapses.²² Potential risk factors for this include LVI, choriocarcinoma or embryonal histology in the primary tumor, and prior groin surgery.^{23, 24} Although prior studies have shown higher rates of pelvic metastases in patients with prior ipsilateral groin surgery, we did not find that this was a predictor of extra-retroperitoneal disease recurrence after pPRLND, with no relapses among 19 patients in our cohort with history of such surgery.^{25, 26} It is worth noting that 16 patients in our series had an ipsilateral pelvic lymph node dissection as well. This finding is limited by the small number of events and the inclusion of only patients without pN+ disease but warrants further study as it may suggest that prior ipsilateral groin surgery should not exclude patients from pRPLND. Similarly, we found a higher rate of relapse in patients with embryonal histology or LVI; however, this was not statistically significant, and neither was associated with relapse on univariable analysis.

Increasingly, the focus of management for early-stage GCT patients with excellent long-term prognosis has been on facilitating patient-centered decision making and reducing long-term morbidity. Patients may be motivated to avoid not only the toxicity of chemotherapy but also the rigors and uncertainty of surveillance regimens.^27–29 Patients who are found to be node-negative following pRPLND are candidates for reduced surveillance regimens.³⁰ Current NCCN guidelines recommend eight to nine abdominal CTs in the first four years of surveillance for patients with CSI disease.¹ Given the risk of secondary malignancy from ionizing radiation, efforts have been made to reduce CT burden in the treatment of testicular cancer by using less intensive imaging schedules or magnetic resonance imaging (MRI).^31–33 In our cohort most relapses occurred within six months, supporting current NCCN guidelines which allow for reduced frequency of surveillance following year one after surgery.¹ Similarly, most relapses occurred in the lungs and were associated with elevated STMs. Surveillance can therefore be narrowed in these patients to consist of only chest x-ray and STM monitoring following one postoperative CT, as is our practice at MSK. The poor sensitivity of plain film chest x-ray for detecting small lung nodules should always been considered in patients at high-risk of developing metastatic disease to the lung. In these patients, prudent use of CT imaging of the chest should be considered. Per NCCN guidelines, patients who are pN0 or pN1 with teratoma only on pRPLND may undergo only one additional abdominal CT scan approximately 3 months after surgery, and then can be followed solely by chest x-ray and STM monitoring.¹ It should be noted that while our study demonstrated no retroperitoneal-only relapses, this finding presupposes a bilateral dissection with high lymph node yield performed at a high-volume center. This is in contrast to data from a randomized trial of patients who underwent modified template dissection and subsequently experienced several retroperitoneal or locoregional relapses in the scrotum or groin.¹⁴ A complete resection and high node count has been demonstrated to both improve diagnostic accuracy of pRPLND (with higher node counts increasing the chance of detecting retroperitoneal disease) and to predict relapse-free survival.^{34, 35}

Patients without disease in the retroperitoneum at pRPLND have an excellent long-term prognosis. However, current imaging modalities and STM testing are neither sensitive nor specific enough to detect micro-metastatic disease, potentially resulting in unnecessary surgeries. It is possible that improved diagnostics may serve to improve patient selection for pRPLND by identifying those patients more likely to harbor viable retroperitoneal disease following orchiectomy, and may identify those patients who, regardless of pathologic stage, may be at risk of distant metastatic. Absent these, we were unable to identify any patient or disease characteristics predictive of relapse in patients with pN0 pRPLND. Further studies are needed to determine whether miRNA or other tests have the requisite sensitivity to improve our ability to diagnose occult metastatic disease compared with existing diagnostic modalities.

Conclusion

Two-year RFS for patients found to be pN0 at the time of pRPLND is excellent (93–95%). In this series, recurrences typically occurred in the lung and within 6 months of surgery, suggesting the presence of subclinical distant metastatic disease at the time of pRPLND. Improved diagnostics may further limit the frequency of pRPLND in patients without retroperitoneal disease and identify patients more likely to experience relapse despite pN0 pRPLND.

Figure 1:

Recurrence-free survival for patients with pN0 pathology at primary retroperitoneal lymph node dissection (RPLND)

Table 1.

Demographic and oncologic characteristics of the 346 patients who underwent primary RPLND and had pN0 pathology

	N=346
Age at RPLND	30 (24, 36)
Clinical Stage
CSI	295 (85%)
CSIIA	51 (15%)
Side of index orchiectomy
Bilaterala	3 (0.9%)
Left	168 (49%)
Right	175 (51%)
Prior ipsilateral groin surgery	19 (5.5%)
Index orchiectomy histology
Yolk sac tumor present	226 (65%)
Teratoma present	225 (65%)
Seminoma present	158 (46%)
Embryonal carcinoma present	307 (89%)
Choriocarcinoma present	22 (6.4%)
Secondary somatic malignancy present	4 (1.2%)
LVI at index orchiectomyb	241 (70%)
Elevated AFP pre-orchiectomy (>15 ng/mL)c	136 (53%)
Elevated HCG pre-orchiectomy (>2.2 U/L)c	141 (58%)
Elevated AFP Pre-RPLND (>15 ng/mL)d	1 (0.3%)
Elevated HCG Pre-RPLND (>2.2 U/L)d	2 (0.6%)
Elevated LDH Pre-RPLND (>246 U/L)e	19 (5.5%)

Abbreviations: AFP, alpha-fetoprotein; CS, clinical stage; HCG, human chorionic gonadotropin; LDH, lactate dehydrogenase; LVI, lymphovascular invasion; RPLND, retroperitoneal lymph node dissection.

^aThree patients had synchronous bilateral disease.

^bLVI not available in 2 patients.

^cPre-orchiectomy HCG not available for 103 patients and pre-orchiectomy AFP was not available for 90 patients.

^dAFP was 26.6 and stable prior to surgery; HCG values were 2.3, and 3.0.

^eLDH not available in 2 patients.

Table 2.

Operative and complication details of 346 patients who underwent primary RPLND and were pN0

	N=346
Nerve-Sparing RPLNDa	293 (85%)
Concomitant ipsilateral pelvic lymph node dissection	16 (4.6%)
Length of Stay	7 (6, 8)
Estimated Blood Loss (mL) at RPLND (N=345)	200 (150, 250)
Total Nodes Removed at RPLND (N=340)	50 (36, 64)
Highest Complication Grade within 30 daysb
No complication	297 (87%)
Grade 1 Complications	20 (6%)
Grade 2 Complications	14 (4%)
Grade 3 Complications	10 (3%)

Abbreviations: IQR, interquartile range; RPLND, retroperitoneal lymph node dissection.

^aDetails on nerve-sparing approach not available in 11 patients.

^bComplication data not available for 5 patients.

Table 4:

Cox univariate regression analysis of factors associated with relapse following pN0 primary RPLND

Predictor	HR	95% CI	p-value
Age at RPLND (per 5-year increase)	1.19	0.95, 1.49	0.13
Clinical Stage
CSI	Ref.	.	.
CSIIA	0.60	0.14, 2.54	0.5
Presence of embryonal carcinoma at orchiectomy	1.34	0.31, 5.70	0.7
Presence of LVI at orchiectomy (N=344)	4.22	0.99, 18.05	0.052
Elevated STMs prior to orchiectomy (N=244)	2.15	0.63, 7.38	0.2
Time from last scan to RPLND (days)	1.00	0.97, 1.03	0.9

Abbreviations: CS, clinical stage; LVI, lymphovascular invasion; RPLND, retroperitoneal lymph node dissection; STM, serum tumor marker.

Clinical Practice Points.

• Patients who undergo primary retroperitoneal lymph node dissection and are found to have no disease in the retroperitoneum face excellent long-term outcomes with a 93% 2-year recurrence free survival.

• Patients undergoing pRPLND for CSI and CSII disease were pN0 in 66% and 28% of cases, respectively, reflecting the limitations of contemporary diagnostics

• Patients who have no disease in the retroperitoneum at primary retroperitoneal lymph node dissection recur most commonly in the lungs or with positive markers and are candidates for reduced surveillance schedules

Highlights.

• Patients without retroperitoneal disease at pRPLND have excellent long-term outcomes

• Recurrences occur early in the lungs, likely due to pre-existing distant disease

• Reduced surveillance is possible for patients without disease at pRPLND

• More sensitive diagnostics are needed to reduce rates of negative pRPLND

Data Statement

Data will not be available as the data set is proprietary to MSK and contains protected health information at the patient-level.