Narrative Review of Developing new Biomarkers for Decision Making in Advanced Testis Cancer

Lucia Nappi; Craig Nichols; Christian Kollmannsberger


Transl Androl Urol. 2021;10(10):4075-4084. 

In This Article


New Biomarkers in GCT: miRNAs

miRNAs are small, non-coding RNAs of about 23 nucleotides involved in post-transcriptional gene regulation thus playing an essential role in many biological processes such as cell differentiation, apoptosis and tumor development.[31,32] Two clusters of miRNAs, miR-371a-3p (miR371) and miR-302/367, have been identified in embryonal stem cells where they control the expression of cell cycle regulating genes.[33,34] Voorhoeve and colleagues reported that in embryonic stem cells micro-RNAs play a pivotal role in the crossover between cell cycle control, pluripotency and chemo-sensitivity.[35] These two clusters have also been described to be overexpressed in GCT tissue and cancer cell lines of both seminoma and non-seminoma while their expression is significantly lower in both benign testicular tissue and teratoma, demonstrating a high specificity for gem cell malignancy.[36] miRNAs are usually secreted in the blood in multi-protein complexes that provide protection from RNAses degradation, increasing their stability in human body fluids.[37] miR371 and miR302/367 clusters were identified in the serum of a pediatric patient with yolk sac tumor in 2011 by Dr. Murray's group for the first time.[38] From that moment on, several studies have been published to demonstrate the detectability of those miRNAs in the blood of patients with GCT.

miRNAs in the Pre-orchiectomy Setting. The first studies were conducted as proof of principles in patients with testicular masses prior to and post- orchiectomy. As reported by Belge first and confirmed by Gillis et al., serum miR-371/372/373/367 levels were significantly higher in the GCT samples than in the control samples.[39] Moreover, these miRNA serum levels rapidly returned to baseline after orchiectomy in patients with CSI disease.[40,41] Those results were independently confirmed in a larger retrospective cohort of 250 patients with CSI GCT by van Agthoven et al. In this study, the test was performed prior to the orchiectomy and revealed a sensitivity of 90%, a specificity of 86%, a positive predictive value (PPV) of 94% and negative predictive value (NPV) of 79% as well as an Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) of 0.95 for miR371.[42]

In their retrospective study, Spiekermann et al. demonstrated increased serum levels of miR371 in germ cell neoplasia in situ, stage I and metastatic GCT, compared with non-GCT control patients. Moreover, miR371 levels were significantly higher in the blood from the tumor ipsilateral testicular vein compared with the peripheral blood, confirming the high specificity of miR371 for GCT.[43]

The largest retrospective study in patients with CSI pre and post orchiectomy was published by Dieckmann et al..[44] In the development cohort of this study, fifty patients with CSI GCT were evaluated for miR371 prior to and post-orchiectomy. miR371 was compared to miR371 tested contextually to miR372, miR373 and miR367 and miR371 was confirmed to have the highest sensitivity and specificity with minimal additional benefit from combination with the other three miRNAs. The sensitivity and specificity of miR371 in the main study (n=160) were 92% and 80%, respectively. The study also demonstrated a direct correlation between GCT stage and levels of miR371 expression.

Those results were confirmed prospectively in a large study published by the same group in 2019.[45] Of the 616 patients evaluated, 522 had CSI GCT and miR371 was evaluated prior to the orchiectomy. In this setting the sensitivity of miR371 was 90.1%, the specificity 94.0%, the PPV 97.2%, and the AUC of the ROC 0.966.

Patients With Metastatic Disease. Several studies have confirmed the high accuracy of circulating miR371 in detecting metastatic disease. In a small retrospective study in pediatric patients with GCT, Murray et al. demonstrated that miR371 and miR302/367 clusters are detectable in the cerebrospinal fluid as well as the serum of these patients with a sensitivity and specificity higher than the classic tumor markers.[46]

In their retrospective study, Dieckmann et al. evaluated miR371 in 49 patients with metastatic GCT prior to, during and after chemotherapy.[44] The miR371 levels decreased during chemotherapy but no data about detectable miR371 post-chemotherapy and presence of active germ cell malignancy (aGCM) were reported.

In 2019, two prospective studies were reported. Our group validated miR371 across the spectrum of GCTs.[47] In this study 132 samples from 111 patients were assigned to three groups on the base of the risk of harboring aGCM. The low risk group (5–25% risk) included patients with post-orchiectomy CSI seminoma and CSIA (negative lymphovascular invasion) non-seminoma with no suspicious signs of relapse on surveillance and patients with or without residual radiologic disease after chemotherapy and normal TM; the moderate risk group (25–50% risk) included patients with post-orchiectomy CSIB (positive lymphovascular invasion) non-seminoma with no suspicious findings of relapse, patients with CSI seminoma and non-seminoma with clinical signs of suspicious relapse on surveillance, and patients with low positive TM after chemotherapy; the high risk group (90–100% risk) included patients with gross clinical metastatic GCTs before starting chemotherapy, patients with testicular mass before orchiectomy, patients with CSIS and patients with obvious viable (high positive TM) residual disease after chemotherapy. Overall, with a median follow up of 15 months, specificity, and sensitivity of miR371 were 100% and 96%, the positive predictive value and negative predictive value were 100%, and 98%, respectively. The AUC of miR371 was 0.96, therefore significantly higher than the CT scan and classic tumor markers in all the clinical situations analyzed.

Another prospective study, analyzed miR371 in 118 patients with metastatic GCT, including 46 patients evaluated at the time of tumor relapse.[45] The investigators confirmed that miR371 expression levels were associated with clinical stage, primary tumor size and response to the treatment. MiR371 median expression was higher in the relapsed patients than in controls and it was overexpressed in 38 of the 46 patients with GCT at the time of the clinically detectable relapse. No data about long-term follow-up were reported in this study.

Equivocal Clinical Scenarios and miRNAs Utility

There are several situations where the suboptimal accuracy of the current diagnostic tests is highly problematic. These include stage IIA with either low positive or TM negative disease and patients presenting with post-chemotherapy residual disease.

Stage IIA Patients. In absence of clearly elevated TM, the management of patients with <2 cm retroperitoneal lymph nodes is challenging due to the high risk of falsely positive enlarged LN.[48] In seminoma, enlarged LN can be related to inflammatory or granulomatous disease, known to have a certain association with this type of GCT.[49] In non-seminoma, retrospective surgical series report that up to 40% of patients with <2 cm retroperitoneal LN have no cancer and are potentially exposed to unnecessary treatments.[50] Based on those data, a biomarker able to discriminate between absence and presence of aGCM is critical.

The operating characteristics of miR371 in patients with CSI and stage IIA have been evaluate by Lafin et al. in 24 patients with seminoma and non-seminoma treated with primary RPLND.[51] The serum levels of miR371 were correlated with the presence or absence of active GCT in the pathology specimens. Of the 24 patients analyzed, 11 (45.8%) presented aGCM, 10 (41.7%) benign tissue and 3 (12.5%) teratoma. miR371 demonstrated high sensitivity (100%) and specificity (92%) to detect aGCM, with an AUC of the ROC of 0.96 in this very equivocal clinical scenario.

The results in moderate risk patients evaluated in our study also confirmed the high accuracy and clinical validity of miR371 in patients with equivocal GCT evidence.[47] In the 46 patients with moderate risk analyzed, miR371 showed a sensitivity, specificity, NPV and PPV of 92%, 100%, 97% and 100% (none of the patients had a false positive result), respectively. The operating characteristics of miR371 in identifying aGCM outperformed CT imaging and TM in this diagnostically very problematic group of patients with an AUC of the ROC of 0.89.

Post-chemotherapy Residual Disease. The management of post-chemotherapy residual disease depends on histology and size of disease. In seminoma, the risk of viable seminoma is low if the residual disease is ≤3 cm. In lesions >3 cm the risk of viable residual seminoma is about 20–30%.[52] FDG-PET has high negative predictive value but the positive predictive value is only around 30%.[21,22,53] In addition, TM are commonly negative in seminoma and do not contribute to characterize the presence of aGCM. In non-seminoma, surgery is recommended for every patient with residual lesions >1 cm.[54,55] However, the analysis of retrospective large surgical series in this patients population have revealed that only 10% of patients have residual aGCM, while 40–50% of the patients have teratoma and 40–50% have necrosis/fibrosis only.[56] Overtreatment is therefore implied in the significant group of patients with non-seminoma without post-chemotherapy residual teratoma or viable GCT who were cured with chemotherapy alone but who cannot currently be differentiated from patients with teratoma or aGCM on the base of the imaging studies.

Several clinical nomograms have been proposed to identify the patients with necrosis/scar tissue only residual disease post-chemotherapy.[57–59] However, the accuracy of those strategies is highly suboptimal and cannot and should not be routinely used for management decisions.

In this setting, a study reported by Leao et al. has demonstrated that miR371 is able to detect the presence of aGCM also in patients with post-chemotherapy residual disease.[60] The study was conducted in 82 GCT patients with residual disease after chemotherapy and prior to RPLND and the serum miRNAs expression was correlated with the pathology results of the RPLND specimens. With the limitation of having only 12 patients harboring aGCM after chemotherapy, miR371 was able to predict the presence of aGCM with a specificity of 50% and sensitivity of 100%. The AUC of miR371 alone was 0.874, comparable to the combined miRNAs 371, 372 and 373 (AUC 0.885). As previously reported, miR371 was undetectable in patients with residual teratoma or fibrosis.

Serum miR371/373/367 were also analyzed by Rosas- post-chemotherapy in a cohort of 109 patients.[61] The median level of these miRNAs was higher in patients who developed a relapse (n=34) than in those who had complete durable remission (n=60). Moreover, serum miR367 was higher in cisplatin refractory patients (n=15) compared to the patients who achieved a complete response.

As demonstrated for primary testicular masses and metastatic disease,[44,45,62] it is very likely that the sensitivity of miR371 in detecting aGCM also in the post-chemotherapy setting relies on the residual aGCM tumor burden after chemotherapy.

Other Equivocal Scenarios. Considering the remarkably high specificity of miR371 in identifying aGCM, it could be clinically valuable to detect GCT in patients with metastatic primary unknown tumors. Especially in absence of elevated serum tumor markers and/or obvious testicular masses, the diagnosis of GCT maybe challenging and requiring invasive procedure to collect tissue for pathology diagnosis. miR371 could be used in this very equivocal cases to identify aGCM and therefore to offer curative treatments.

Limitations of miR371. In order to promote the scalability and implementation of miR371, standard methodology should be used for miRNAs extraction and expression analysis. Normalization of miR371 using internal housekeeping miRNAs is recommended to reduce the variability of results related to technical issues. Moreover, hemolysis should be always considered as it could affect the miRNAs expression results.[46,63] Although miR371 is potentially useful to predict the presence of aGCM, in the post-chemotherapy setting its utility is limited by the negative expression in teratoma. As a consequence, while patients with positive miR371 have high risk of having aGCM in case of negative serum/plasma miR371 the presence of teratoma cannot be rule out. Therefore, several efforts have been made to identify biomarkers specific for teratoma to be integrated with miR371 to characterize the pathology composition of residual post-chemotherapy masses.

Potential Teratoma Biomarkers: miR375-3p. miR375-3p (miR375) is a miRNA overexpressed in teratoma and yolk sac tumor tissue.[64] Unlike miR371, this miRNA is not specific only for teratoma or GCT but low levels of miR375 are detectable in healthy volunteer (no cancer patients)[65] and it is overexpressed also in other solid malignancies.[66] The detectability of miR375 in the blood of patients with GCT has been reported in 4 studies but with controversial results.

Belge et al. reported about serum miR375 in 21 patients with teratoma, 12 patients with other GCT, and 12 control samples.[67] The levels of miR375 were not different between teratoma and other GCT patients or controls. The AUC of the ROC for miR375 in identifying teratoma was 0.524.

The serum miR375 was retrospectively analyzed by Lobo et al. in a cohort 36 patients undergoing chemotherapy followed by RPLND at different time points: pre-chemotherapy, post-chemotherapy, pre-RPLND and post-RPLND.[68] 15/36 patients presented teratoma and no significant differences were found between them and the patients with benign or residual aGCM prior to or post-chemotherapy or post-RPLND.

Kenigsberg et al. recently reported about the serum expression of miRNA-375 and miR375-5p prospectively evaluated in 40 GCT patients (19 teratoma, 18 benign pathology, 2 viable GCT, 1 embryonal rhabdomyosarcoma) who underwent post-chemotherapy RPLND.[69] miR-375-5p was undetectable in all samples examined. The AUC of the ROC for miR375 was 0.503. Of the 19 patients with teratoma, 16 had elevated miRNA-375 with a sensitivity of 84.2%; the specificity was 35.3%. PPV and NPV were 59.3% and 66.7%, respectively.

Our group proposed an integrated analysis of miR371 and miR375 to predict the risk of aGCM and teratoma in patients with post-chemotherapy residual disease. In our study, plasma miR375 and miR371 were analyzed in 41 patients with confirmed pure teratoma and compared with patients with either no tumor or pure classic seminoma (n=59).[70] Overall, sensitivity and specificity of miR375 in identifying teratoma were 78% and 80%, respectively; the AUC was 0.7. However, the integrated analysis of both miR371 and miR375 increased the discriminative power to distinguish benign to teratoma and aGCM tissue in patients with residual disease post-chemotherapy.[70] Although small and preliminary, our data suggest that there is a potential in developing an integrated clinical-molecular model to predict teratoma, aGCM and benign tissue in patients with post-chemotherapy residual disease that may increase the accuracy of the current diagnostic tools.

Methodologic differences among the above mentioned studies may contribute to explain the difference in the results. The collecting tubes were different. While Nappi et al. used cell-free DNA BCT (Streck tubes) and plasma for their analysis, in the other three studies the blood was collected in serum separator tubes and miR375 was analyzed in serum. Although no data about accuracy detection of miR375 in cell-free DNA BCT tubes vs. serum separator tubes are available, some evidence suggest that preservation solutions for cell-free DNA designed to prevent cell lysis and apoptosis also minimize release of cellular miRNAs, reducing the unwanted background after blood collection and improving the detection of the targeted miRNAs of interest.[71]

Moreover, while Nappi et al. demonstrated a clear correlation between the size of the teratoma component and the plasma expression of miR375, no data regarding the correlation between the size of the teratoma in the post-chemotherapy residual disease and miR375 level of expression were provided in the other studies. This is relevant and may explain the differences of the results observed in the above mentioned studies. Furthermore, the number of the teratoma patients included in each study was different: in the Belge publication only 5 patients with teratoma had CS II-III while the majority of patients had CSI disease. In Lobo's paper, it is unclear from the manuscript whether or not the teratoma patients had only pure teratoma or included also patients with mixed histology.

Status of miRNAs Validation in GCT

Although several independent groups have demonstrated the high accuracy of miR371 in detecting aGCM, there is still the need to further validate this miRNA in larger prospective trials before introducing its utilization in the clinical practice. While its clinical validity has been demonstrated in patients with obvious disease (pre-orchiectomy and in patients with unequivocal metastatic disease) the operating characteristics of this miRNA in patients with more equivocal disease or microscopic disease are still unknown.

In order to validate the clinical utility of miR371 in the diagnosis and follow up of GCT patients and to define the cut off of sensitivity to detect GCT, two prospective clinical trials have been designed and are currently enrolling patients.

The Children's Oncology Group AGCT 1523 trial (Active Surveillance, Bleomycin, Carboplatin, Etoposide, or Cisplatin in Treating Pediatric and Adult Patients With Germ Cell Tumors) is a multi-arms study designed in the pediatric population with an optional biomarker arm enrolling also adult GCT patients to evaluate miR371/miR372/miR373 and miR367 clusters during the surveillance of CSI patients.[72]

In adult GCT patients, S1823, a SWOG international cohort clinical trial, has been activated in June 2020 and is currently recruiting GCT patients.[73] This study primary objective is to correlate miR371 to the presence of relapse in patients with CSI and CSIIA on surveillance; the primary endpoint is to establish the PPV of miR371 in detecting GCT relapse. Since the study plans serial blood tests, it will be possible to establish if and when exactly miR371 becomes positive during surveillance and to correlate the miR371 status with the clinical presentation of the disease at the time of tumor relapse.

Future Prospective

In addition to S1823, another clinical trial has been planned to validate the use of miR371 expression in patients with GCT with <3 cm retroperitoneal LN to select patients for either surgery or observation. Moreover, there is a lot of enthusiasm in the GCT community for joint efforts to build up integrated clinical and biomarkers models for the identification of teratoma and benign residual disease in patients with non-seminoma presenting with post-chemotherapy residual disease. In this setting, radiomics has showed promising results and need to be further explored.[74,75]