Introduction

Approximately 30% of patients with renal cell carcinoma (RCC) were diagnosed with high-risk localized disease. Among this group, over one-third experienced recurrence after undergoing nephrectomy.1 2 The use of perioperative therapy with targeted molecular therapy or immune checkpoint inhibitors (ICIs) was controversial. Recently, adjuvant pembrolizumab and sunitinib show the ability to delay recurrence in KEYNOTE-564 and S-TRAC trials.3 4 In the S-TRAC trial, no significant benefit in overall survival (OS) was observed (HR 0.92, 95% CI 0.66 to 1.28, p=0.6). However, in the KEYNOTE-564 trial, pembrolizumab demonstrated a statistically significant improvement in OS compared with placebo (HR 0.62, 95% CI 0.44 to 0.87, p=0.0024) over a median follow-up period of approximately 57 months.5

Nephrectomy is typically the primary treatment for localized RCC, and systemic therapy has demonstrated potential in selected cases. Preoperative systemic therapy has several benefits, including the potential to convert an unresectable tumor into a resectable one, downstage tumor thrombus (TT) to make surgery less challenging, and enhance renal function preservation when used alongside a PN. Overall, the systemic therapeutic regimen has the potential to effectively reduce the size of the primary tumor, which is crucial for managing difficult cases of locally advanced RCC.5

The advances in the development of systemic regimens have led to an OS improvement in patients with metastatic RCC.6 The promising results from five trials have had a major influence on the landscape of advanced RCC.7–11 Axitinib plus ICI regimens showed high objective response rate (ORR) of 60.4% and 55.9% in KEYNOTE-426 and Javelin renal 101 trials, respectively.8 9 The high ORR observed in this study may have potential implications for locally advanced RCC by facilitating tumor shrinkage, which might improve the feasibility of complex nephrectomy. On the other hand, the neoadjuvant treatment with ICI depends on the hypothesis that the primary tumor may serve as an antigen source, triggering a specific and long-lasting immune response against cancer. Neoadjuvant immunotherapy has demonstrated a more sustained response compared with adjuvant immunotherapy in the context of melanoma.12 13 Limited information is available regarding the safety and efficacy of ICIs and targeted molecular therapy coadministered prior to nephrectomy and how drug administration or immune-related adverse events (AEs) impact on surgery.

Toripalimab, a humanized antiprogrammed cell death protein 1 (PD-1) IgG4 antibody,14 is approved in China for the treatment of six cancer indications, including melanoma, urothelial cancer, non-small-cell lung cancer, and esophageal squamous cell cancer. The efficacy of the toripalimab and axitinib treatment regimen is currently being assessed in a phase III trial RENOTORCH (NCT04394975) conducted in China. Promising results regarding progression-free survival (PFS) (18.0 months) were reported in October 2023.15 This study is a phase II clinical trial performed at a single center, focusing on the safety, efficacy, and biomarkers of the combination of toripalimab and axitinib in locally advanced clear cell RCC.

Patients and methodsStudy design

This study was conducted as a phase II trial, with a non-randomized, open-label, and single-center design. The study registered on clinicaltrials.gov (NCT04118855) in accordance with the principles of the Declaration of Helsinki.16 Following informed consent and baseline screening, the participants received a 12-week treatment of axitinib plus toripalimab before undergoing curative nephrectomy.

Study population

The main inclusion criteria consisted of the following: non-metastatic disease (T2-T3N0-1M0), clear cell RCC confirmed by biopsy, scheduled for either PN or RN as part of their treatment plan; Eastern Cooperative Oncology Group performance status of 0 or 1, adequate organ and marrow function as determined by specific laboratory tests specified in the study, and willingness to adhere to scheduled visits, treatment plans, laboratory tests, and other study procedures. The main exclusion criteria encompassed the presence of any other cancers within the past 5 years, the use of other systemic antitumor treatments (including targeted therapy and immunotherapy), the requirement for immediate nephrectomy to alleviate symptoms, current use of immunosuppressive agents, pregnancy or breastfeeding, and a history of autoimmune disease or syndrome.

Treatment, surgery approach, and follow-up

Axitinib was orally administered at a dose of 5 mg twice daily for a duration of 12 weeks prior to the scheduled nephrectomy. The dose reduction of axitinib was performed according to the manufacturer’s instructions. Toripalimab was administered intravenously at a fixed dosage of 240 mg on day 1 of a 3-week, 6-week, and 9-week cycle, for a maximum of three doses. After completing the 12-week treatment, the administration of axitinib was stopped for a period of 3 days, after which a surgical procedure was performed. The surgeon and patient would decide together which procedure to be performed depending on the post-treatment imaging assessment. Following successful completion of neoadjuvant therapy, patients underwent PN or RN (open, laparoscopic, or robotic at the surgeon’s discretion). All patients undergoing radical surgery will receive lymph node dissection. Primary tumor specimens were immediately sent to pathology for further processing. For pathological complete response (pCR) patients, all tumor tissues are sliced, and no residual tumor cells are confirmed under the microscope by an experienced pathologist, Dr Wang (ZW). The clinical follow-up, as required by the protocol, occurred 1 month after the surgery and then every 3 months for the next 5 years.

Endpoint and outcome assessment

The primary endpoint is ORR assessed using the Response Evaluation Criteria in Solid Tumors, which categorizes the response as CR or partial response (PR).17 Enhanced MRI (prior) or enhanced CT imaging evaluations were performed at baseline, week 7 and week 12 following the initiation of the therapeutic intervention. Tumor size, level, and length of TT were assessed by an experienced radiologist, Dr Wu (GW), who has over 13 years of experience in renal imaging. Morphological measurement was performed using contrast-enhanced imaging. Tumor size was measured via determining the largest cross-section of the tumor in the axial image. The level and length of the TT were assessed in the coronal image. The three-dimensional volume of interest of the TT was manually delineated slice by slice using ITK-SNAP software (V.3.6.0). The extent of inferior vena cava involvement was categorized based on the Mayo staging system’s classification of TT level.18 For apparent diffusion coefficient (ADC) maps, a volumetric radiomic analysis was performed to measure the mean value and entropy, which represents the randomness of gray-level distribution described by Haralick et al to reflect change in pathology and immunology through imaging modality.19 20

The secondary endpoints include disease-free survival (DFS), safety, and perioperative complication rate. Patients were monitored for AEs from the initiation of treatment until the completion of the study. The toxicity assessments followed the guidelines specified in the Common Terminology Criteria for Adverse Events V.5.0 (CTCAE V.5.0).

Biomarker analysis

To develop the predictive model for patent selection and the change of immune microenvironment before and after neoadjuvant therapy, NanoString nCounter PanCancer IO 360 Panel and multiplexed immunofluorescence (mIF) staining were performed as described in online supplemental methods.

Statistical considerations

The primary endpoint was examined using a Bayesian design. Here, we assumed that the distribution of πORR follows a β distribution with parameters (0.64–1.36), yielding a mean of 0.32 and a SD of 0.27. πORR-β (0.64–1.36) with an average value of 0.32. It can be inferred that the efficacy of axitinib plus toripalimab in patient treatment would be deemed satisfactory if the posterior probability of πORR exceeding 0.32 is ≥80%. The observed proportion of responses was used to calculate 95% credible intervals (CrI). The secondary endpoint was used to calculate 95% CI.

Statistical analysis for survival and biomarkers was performed using SPSS V.26.0 and GraphPad Prism 8. The Reverse Kaplan-Meier method was employed to calculate the median follow-up time. Pearson’s χ2 test or Fisher’s exact test was used to compare categorical variables. The Mann-Whithey U test or unpaired t-test was used to compare continuous variables. Statistical significance was set at p<0.05.

ResultsPatient characteristics

From July 2020 to August 2022, 20 patients, with a median age of 61.5 years (range: 42–78), were enrolled and treated with axitinib plus toripalimab (figure 1). During the same period, 161 patients with T2-3N0-1M0 RCC underwent surgery in our center. The characteristics of patients were summarized in table 1. 17 patients were male (85%), and 15 patients demonstrated TT in baseline.

Figure 1

Trial design and profile. (A) Study schema. (B) Trial profile. Twice a day.

Table 1

Baseline patient characteristics

Radiological and pathological response

A total of 20 patients underwent evaluation for radiological response, with pretreatment and post-treatment MRI or enhanced CT scans being paired. PR was observed in 9 out of 20 patients, resulting in an ORR of 45% (95% CrI: 23.2–66.8) as depicted in figure 2A. The posterior distribution of πORR had a mean of 0.44 (95% CrI: 0.24–0.64), estimated within the range of 9.64–12.36 from our analysis, meeting the predefined primary endpoint with an ORR of 32%. Results in figure 2B indicated that 19 out of 20 patients showed tumor shrinkage. Median maximal reduction of primary renal tumor diameter was 26.7%, with a range of −2.0% to 40.3% (online supplemental table 1). TT response was shown in online supplemental figure 1 and table 1; the median TT length exhibited a numerical reduction from 9.1 cm (range: 2.4–25.8) to 8.2 cm (range: 1.7–23.4), while median TT volume exhibited a reduction from 29.3 cm3 (range: 2.5–318.4) to 15.1 cm3 (range: 1.0–273.6). Three patients experienced TT enlargement. Other patients exhibited no change and decrease in TT level. No disease progression was observed during the administration of axitinib in combination with toripalimab.

Figure 2

Treatment response. (A) Spider plot of tumor response evaluated by radiological imaging. (B) Waterfall plot of tumor response. (C) The change between clinical T stage and pathological T stage.

As shown in figure 3C, pathological downstaging was observed in 8 out of 19 patients. Of note, four patients with radiological PR were found no evidence of disease (pT0) in final surgical pathological staging (online supplemental table 2).

Figure 3

Treatment follow-up. (A) Swimmer plot. (B) Disease-free survival.

Drug treatment and adverse events

As shown in online supplemental table 3, no grade 4 or 5 AEs were observed in this trial. Hypertension (35%), dysphonia (30%), stomatitis (25%), fatigue (15%), hand-foot syndrome (15%), pruritus (15%), and urinary tract infection (15%) were the prevailing AEs among the patients.

A total of 15 out of 19 patients underwent surgery as scheduled, while 2 of them experienced drug discontinuation or reduction due to drug-related AEs and underwent early surgery. One patient missed surgery due to the patient’s declination. One patient had a delay of surgery after 12 weeks of treatment due to grade 3 hyperglycemia. A patient discontinued the use of drugs and received hydrocortisone 100 mg intravenously due to grade 3 elevated levels of AST/ALT after 7 weeks of treatment and then underwent upfront surgery. A patient reduced axitinib dose to 5 mg orally once a day due to the presence of proteinuria and requested early surgery. One patient declined future treatment after 6 weeks and requested early surgery. All other patients continued their axitinib dose 5 mg orally twice a day. One patient experienced interruption of axitinib and toripalimab treatment due to COVID-19.

Surgery, pathological response, and postoperative complications

As shown in table 2, a total of 19 patients underwent surgery (1 PN and 18 RN). Only one patient with T2a disease with strong willingness to persevere renal function underwent PN, all the other patients selected RN. 14 patients underwent open surgery, while the remaining 5 patients underwent laparoscopic or robotic surgery. The median operative time was 160 min (range: 80–420). The median estimated blood loss was 300 mL (range: 40–2500). Adhesion to adjacent tissues was observed in 36.8% of patients (7 out of 19) during the surgical procedure. Among these patients, four exhibited severe desmoplasia. However, no unexpected or limiting consequences were observed. The study found that the median duration of hospitalization after surgery was 4 days (range: 2–20). Two patients underwent intraoperative complication (table 2). The incidence rate of all grade and grade ≥3 postoperative complications was 42.1% (n=8) and 5.3% (n=1), respectively.

Table 2

Surgical parameters and perioperative complications in 20 patients receiving treatment

Survival and recurrence

All patients did not receive adjuvant therapy after surgery in this study. With a median follow-up of 21.3 months (range 5.9–35.1 months), four patients experienced recurrence during the follow-up (figure 3A). The recurrence pattern and sites were shown in online supplemental table 4. The median DFS was not reached. The estimated DFS rates at 1 year and 2 years were 84.7% (95% CI, 68.7% to 100%) and 84.7% (95% CI, 68.7% to 100%), respectively (figure 3B). We did not observe a difference in DFS, according to PR or pCR status (online supplemental figure 2). Only one patient died because of the disease progression (lung and liver metastasis).

Characteristics of MRI and biomarkers before or after treatment

For clinical outcome, entropy at baseline (4.76±0.78 for PR and 4.18±0.62 for SD) and difference between baseline and post-treatment (−25.9±21.8% for PR and −6.8±13.0% for SD) were not significant between PR and SD groups (p=0.069 for both sets, figure 4A). The entropy values at post-treatment were 3.41±0.69 for PR and 3.86±0.61 for SD. The ADC values at baseline were 1.35±0.29, *10–3 mm2/s, for PR and 1.54±0.52, *10–3 mm2/s, for SD, and the ADC values at post-treatment were 1.24±0.34, *10–3 mm2/s, for PR and 1.43±0.43, *10–3 mm2/s, for SD. The difference between baseline and post-treatment for entropy and ADC did not show statistical significance (p=0.129, 0.464, 0.193, and 0.917, respectively) (figure 4B).

Figure 4

(A and B) MRI and pathological alterations in patients before and after axitinib plus toripalimab administration. (A) Example of highly immune infiltrated tumor; (B) example of minimally infiltrated tumor. ADC: apparent diffusion coefficient. (C) Baseline biomarkers correlated with response to treatment analyzed by NanoString IO 360. The volcano plot displays differential gene expression at baseline. (D). Relative fraction of tumor-infiltrating lymphocytes. Each row represents an immune cell subpopulation estimated by multiplexed immunofluorescence, including the density of PD1+ cells, PD1+CD8+ cells, PD-L1+ cells, PD-L1+CK+ cells, CD3+ cells, CD4+ cells, CD8+ cells, CD20+ cells, CD56+ cells, FOXP3+ cells, CD68+CD163- macrophage cells M1, CD68+CD163+ macrophage cells M2. The relative infiltrations of each lymphocyte subpopulations were normalized into a z-score. (E). Comparison of density levels of tumor-infiltrating PD1+ cells, PD1+CD8+ cells, PD-L1+ cells, PD-L1+CK+ cells between partial response (PR) and SD patients. *, p<0.05. (F). Comparison of density levels of tumor-infiltrating CD3+ cells, CD4+ cells, CD8+ cells, CD20+ cells, CD56+ cells, FOXP3+ cells, M1 macrophages, and M2 macrophages between PR and SD patients. No significant differences were found. (G). Representative multiplexed immunofluorescence images showing the simultaneous staining of panel 1 (left): DAPI (blue), CD8+ (red), PD-1+ (green), PD-L1+ (yellow), and CK (gray). Panel 2 (right): DAPI (blue), CD3+ (red), CD4+ (green), FOXP3+ (yellow), CD20+ (orange), and CK (gray) in PR (up) and SD (bottom) patients. Scale bar: 20 µm.

We evaluated the biomarker changes of tumor tissues in patients prior to treatment (baseline). To enhance the sensitivity of biomarkers, the patients with moderate response (tumor shrinkage range from >15% to <30%) were removed from RNA and mIF analysis. The NanoString IO 360 results indicated that multiple immune factors, for example, ARG1, MAGEs, and DC, had a lower baseline level in PR patients compared with SD patients. Transforming growth factor-beta and major histocompatibility complex 2 (MHC2) had a higher baseline level in PR compared with SD patients (statistically insignificant). Differential gene expression is shown in figure 4C. The expression of three genes, namely, TNFSF13, APLNR, and CCL3-L1, was significantly higher in PR patients compared with SD patients. Tumor tissues obtained prior to treatment were subjected to tumor microenvironment analysis in order to better understand the difference between PR and SD patients. The density of immune cells was analyzed using mIF. Tumor-infiltrating immune cell subpopulations exhibited a significant difference between the PR patients and SD patients (figure 4D). Higher densities of PD-1+ cells, PD-L1+ cells, PD-1+CD8+ cells, and M1 macrophages were found in the PR group than in the SD group (figure 4E–G), suggesting a highly immune-infiltrated microenvironment with a high percentage of PD-1+CD8+ exhausted T cells in PR patients prior to treatment. The results of individual levels were presented in online supplemental figure 3.

Discussion

Our trial showed that axitinib plus toripalimab can be used in a presurgery setting of locally advanced RCC. We observed promising antitumor activity of this combination with acceptable toxicity. Most patients showed primary tumor and TT shrinkage. The overall toxicity was comparable between this trial and the previous trial involving an axitinib-ICI combination in a presurgery setting.21

A part of non-metastatic RCC was high risk for recurrence after nephrectomy. Clinical utilization of the neoadjuvant approach to treating RCC is limited to the selected cases that are need for downsizing primary tumor or TT to improve the feasibility of nephrectomy. For now, although no study provides evidence for OS improvement in this setting, there is consensus that a complete section of primary tumor can expand clinical benefits in locally advanced and metastatic settings.22 23 The application of neoadjuvant therapy in RCC during the era of tyrosine kinase inhibitor (TKI) treatment has been investigated, but the efficacy of combining neoadjuvant TKI with ICI is not well understood. Here, we reported our study that in a cohort with 75% TT, neoadjuvant therapy with axitinib plus toripalimab showed strong antitumor activity and facilitated following nephrectomy. Moreover, a comparison of the tumor diameter between weeks 7 and 12 revealed ongoing tumor shrinkage, suggesting continued therapeutic effect. Based on these observations, we advocate for a 12-week treatment duration for these patients, which we believe offers superior benefits. And for target therapy, we chose axitinib in our study because of the short half-life of axitinib. Previous studies employing axitinib as a monotherapy yielded an ORR ranging from 17% to 46%. Additionally, the median reduction in diameter of the primary renal tumor was reported to be 28.3% (range 5.3–42.9), which is consistent with our results.24–26 However, the efficacy of ICI alone seems limited in this setting. NeoAvAx study using axitinib plus avelumab regime showed ORR of 30%, and the median primary tumor downsizing was 20% (range: +3.8–43.5).21 Karam et al conducted a trial using sitravatinib plus nivolumab for 6–8 weeks, yielding a lower ORR of 11.8%. However, this regimen demonstrated comparable DFS to our current trial. Notably, our study focused on a high-risk population, with 75% of patients exhibiting TT. The divergent treatment durations, with our trial employing a 12-week regimen compared with Karam et al’s 6–8 weeks, likely contribute to the observed differences in outcomes.27 A study using three cycles of nivolumab showed that most patients had a minor change of primary tumor.28 Consistently, another study using four cycles of nivolumab demonstrated that all patients had stable disease prior to nephrectomy.19 Overall, TKI and TKI plus ICI seem to be more effective than ICI alone in a presurgery setting. Primary tumor response is comparatively moderate as compared with metastatic sites in previous reports, and the results were confirmed by recent post hoc analysis.29–32

A TT response is the most important to reduce the extent of surgery and the associated surgical morbidity.33 34 We observed a significant reduction in TT volume following neoadjuvant therapy. The downstaging of TT levels was observed in 33% of patients, thereby enhancing the possibility of achieving complete resection. Of 15 patients with TT, 1 patient showed TT progression during the drug administration. This result is comparable with the previous study that used axitinib alone.24

An intriguing finding in this study is the identification of four patients (20%) who exhibited a pCR. It is noteworthy to mention that all patients classified as pT0 exhibited a PR. The study conducted by Panian et al also discovered that among a group of 52 patients with advanced or metastatic RCC who received immunotherapy prior to cytoreductive nephrectomy, 13% of them showed pCR.35 pCR was observed in case reports of patients treated with ICI alone, a combination of TKIs and ICI, and a combination of PD-1 inhibitors and cytotoxic T-lymphocyte-associated protein 4 inhibitors36–40 but not in TKI alone treatment. In addition, a case regarding RCC with TT reported a pCR after a 3-month treatment of avelumab plus axitinib.41 Overall, ICI-based combination therapy could induce pCR as compared with TKI alone, which may also lead to better survival.

In our previous study, we observed comparable survival outcomes among patients with TT who underwent neoadjuvant therapy and those who did not.42 Additionally, a study conducted on 72 patients with advanced RCC indicated that there was no significant difference in 3-year PFS between patients who underwent pathological downstaging and those who did not.35 Consistency with these results, we did not observe a significant difference in DFS according to PR or pCR status. However, we noticed that four pCR patients in our study did not suffer from any recurrence at the end of this study, which necessitates further investigation to establish the validity of using pCR as a predictor of survival outcomes.

This study aims to explore the safety of axitinib plus toripalimab. Although the safety of axitinib-ICI combination regimen has been investigated in an advanced setting, our knowledge of its safety in a presurgery setting is limited in NeoAvAx study.21 Their study enrolled 40 locally advanced high-risk patients without the information of TT. A total of four grade 3a and one grade 4a Clavien-Dindo AEs were identified in the 30-day postoperative period.21 In our study, one patient with CKD grade 2 renal function impairment before pre-operative developed uremia within 90 days post-surgery. The patient subsequently recovered after undergoing hemodialysis. In accordance with NeoAvAx study, around a quarter of patients showed drug-related grade 3 AEs induced by axitinib-ICI combination regimen.21 The incidence of AEs is generally lower in a presurgery setting than that in a metastatic setting. This difference in incidence may be attributed to the lower dosage of exposure.7–9 11 The combination of axitinib and toripalimab is generally well tolerated, with only one patient discontinuing treatment due to elevated levels of AST/ALT after two cycles. The delay of surgery is infrequent in this study. The present study only observed two intraoperative complications that were resolved by splenectomy and vessel repairment. There were no readmissions within 30 days that were related to the surgery. The occurrence of Clavien-Dindo complications grade ≥3 suggests that they are less likely to be attributed to immunotherapy and more likely to be associated with the extent of surgery. Localized advanced RCC is at a high risk of progression and requires careful monitoring for disease progression during a neoadjuvant setting.

Exhausted T cells exhibit notable functional impairments in both cytokine production and proliferation.43 The activation of CD8+T cells was noted to transpire at an earlier stage of RCC, subsequently followed by the progression of exhaustion due to persistent antigen stimulation. This phenomenon was found to be linked with unfavorable clinical outcomes.44 PD-1 has been recognized as one of the markers for exhausted T cells, and the inhibition of PD-1 has demonstrated a partial revival of T cell activity.45 The CheckMate-010 trial revealed a positive correlation between high levels of tumor-infiltrating CD8+T cells expressing PD-1 and longer immune-related PFS in patients treated with nivolumab. Conversely, patients with low levels of CD8+PD-1+ cells, who also lacked PD-L1 expression, did not respond to nivolumab.46 Thus, a tumor microenvironment characterized by a significant infiltration of immune cells and a high proportion of CD8+PD-1+ exhausted T cells could serve as a favorable indicator for the effectiveness of PD-1 blockade therapy.

The neoadjuvant therapy followed by adjuvant therapy approach, established in patients with lung cancer47, has not undergone extensive investigation in RCC. The primary challenge lies in the absence of reliable pathological or clinical markers for predicting patient prognosis. In this trial, we evaluated patient response to a TKI-IO regimen and assessed alterations in the immune microenvironment subsequent to neoadjuvant therapy. Individuals exhibiting favorable responses to neoadjuvant treatment and presenting a high risk of recurrence may derive benefit from receiving adjuvant therapy.

The study has limitations resulting from a small sample size of patients, being conducted in a single institution, and having a relatively short follow-up period. A decrease in the rate of accumulation was observed following the outbreak of COVID-19 in Shanghai, which resulted in a significant impact on our outpatient visits. Furthermore, the preference of patients for adjuvant therapy rather than neoadjuvant therapy contributes to the opposition of enrolling subjects, as perceived by patients as well as physicians. The adequacy of a 12-week treatment duration for all patients is uncertain due to the continuous dynamics of tumor response. Also, correlative analyses were not prespecified in the trial protocol, lacked adjustment for multiplicity, and were based solely on changes from baseline without control in this single-arm study, limiting their interpretability.

In conclusion, the results from this trial highlighted the efficacy of axitinib plus toripalimab to downsize primary RCC with an ORR of 45% and strong TT response. Biomarker analysis showed that upregulation of ARG1, MAGEs, DC, TNFSF13, APLNR, and infiltration of exhausted CD8+T cells at baseline were predictive indicators of axitinib plus toripalimab.