Intratumorally injected pro-inflammatory allogeneic dendritic cells as immune enhancers: a first-in-human study in unfavourable risk patients with metastatic renal cell carcinoma

INTUVAX production

INTUVAX was manufactured at the Good Manufacturing Practice (GMP) compliant facility of the Cancer Centre Karolinska (CCK), Stockholm, Sweden. Monocytes were isolated from a leukocyte concentrate from a healthy blood donor at the Blood Centre, Karolinska University Hospital. The monocytes were cultivated in CellGro® DC Medium (CellGenix™), a serum free cell culture medium, supplemented with 100 ng/mL granulocyte macrophage colony-stimulating factor (GM-CSF) and 20 ng/mL interleukin 4 (IL-4) (both cytokines from CellGenix™). Thereafter, the cells were cultured in a closed system and incubated at 37 °C in 5% CO₂ atmosphere. After differentiation, the CellGro® DC Medium was supplemented with 100 ng/mL GM-CSF and 20 ng/mL IL-4 and added to the cell suspension. To induce DC activation/maturation, toll-like receptor (TLR) 7/8 agonist R848 (2.5 μg/mL; InvivoGen), TLR3 agonist Poly I:C (20 μg/mL; Sigma-Aldrich), and human recombinant interferon gamma (IFN-γ; Imukin) (1000 U/mL; Boeringer-Ingelheim) were added to the culture medium. After maturation of DCs, mature pro-inflammatory DCs were harvested and resuspended in heat-inactivated AB plasma obtained from the Blood Centre, Karolinska University Hospital, and supplemented with 10% DMSO. One mL of the cell suspension was filled in each cryovials (CryoTubes™, NUNC) and subsequently deep-frozen to −150 °C, using a computerized gradient freezer (Planer) and thereafter transferred to a − 150 °C freezer. The final product, INTUVAX, was thawed and assessed for release tests (including testing for sterility, mycoplasma and endotoxin) for clinical use. All doses of INTUVAX that were used in this study originate from the same batch.

Patients

Men and women at least 18 years of age with newly diagnosed synchronous mRCC were enrolled. Inclusion required a size of the primary kidney tumor of at least 4 cm in longest diameter, as determined by CT, and at least one measurable distant metastatic lesion. Patients should have an Eastern Cooperative Oncology Group (ECOG) performance status <3 and were required to be candidates for nephrectomy and to have adequate hematological parameters. Therefore, patients with a life expectancy of less than 3 months, brain metastases, active or latent virus disease (HIV, HBV and HCV), other malignancy or ongoing active autoimmune disease, which required treatment with systemic immunosuppressive agents were excluded.

Study design and treatment

The study was a prospective single armed, open label phase I/II study. Patients on the waiting list for nephrectomy were enrolled in 3 cohorts of 3–5 patients per cohort. All patients in the same cohort were treated at the same dose level, i.e. 5 × 10⁶, 10 × 10⁶ or 20 × 10⁶ viable and MHC class II expressing DCs. Two CT-guided injections with 14 ± 3 days of interval were administered into the viable part of the primary renal tumor identified with contrast enhanced CT scan. All INTUVAX doses used in the study were from the same batch.

The planned nephrectomy was performed within 28–42 days after the first vaccination. The patients were hospitalized for 24 h after each vaccination. The patients then returned to the clinic for safety visits one week after each vaccination, in connection with nephrectomy and 3 months after nephrectomy. Monitoring of vital signs, physical examination, safety lab and collection of adverse events (graded according to the National Cancer Institute Common Toxicity Criteria (CTC) version 4.0), were made at visits during the treatment period and at the follow-up visit, 3 months after nephrectomy.

Immunologic parameters including inflammatory immune serology, immune cell parameters, INTUVAX cell tracking and analyses using enzyme-linked immunosorbent spot (ELISpot) were monitored in connection with the vaccinations (samples taken at the day for vaccination and the day after). To evaluate potential auto- and allo-immune events, auto- and allo-immunization parameters were evaluated at screening and at follow-up visit 3 months after nephrectomy. Information on disease progression and survival data were obtained for all patients. Tumor follow-up, according to Response Evaluation Criteria In Solid Tumors (RECIST) 1.1 criteria, was conducted at baseline and at follow-up visit 3 months after nephrectomy. An extended follow-up to evaluate clinical outcome was further performed.

Study oversight

This study was approved by the local ethics committee in Uppsala, Sweden, and was conducted in accordance with Good Clinical Practice guidelines, as defined by the International Conference on Harmonisation. All patients provided written informed consent that was based on the principles of the Declaration of Helsinki. A safety monitoring committee reviewed safety during this study. The study was designed by the authors in collaboration with the sponsor (Immunicum AB, Gothenburg, Sweden).

End points

The primary endpoints were registration of adverse events and changes in vital signs and laboratory parameters from baseline. Secondary end points included evaluation of tumor-specific immune response using immunohistology parameters of the renal tumor post nephrectomy, tumor-specific ELISpot assay, inflammatory immune parameters, systemic cytokine and chemokine production, immune cell distribution and activation, vaccine cell tracking and evaluation of potential auto- and allo-immunization. Finally, tumor response by CT scan was evaluated, according to RECIST 1.1.

Blood sample collection and storage

Heparinized blood was drawn before treatment, during the vaccinations, and at the end of the study. Peripheral blood mononuclear cells (PBMCs) were isolated by centrifugation on Metrizoate-ficoll (Lymphoprep, Nycomed/Axis-Shield PoC/Alere Technologies AS, Oslo, Norway) using a standard method, and thereafter aliquoted and cryopreserved for later use. Serum samples were also collected before treatment and at the end of the study and stored at −80 °C until analysis.

HLA-typing

At screening, evaluation of the patient’s HLA-type (HLA-A, HLA-B and HLA-DR) was made. The samples were analyzed at the local clinical immunology laboratory at site (Dept of Clinical Immunology, Uppsala University Hospital, Uppsala, Sweden). HLA-typing of vaccine cell donor was also performed.

Tracking of injected INTUVAX cells

Flow cytometry analysis for donor cell tracking was performed on peripheral blood samples (taken just before vaccination and 1 h and 20–24 h after each vaccination) after lyzing of the erythrocytes with ammonium chloride and washing using a standard procedure. The samples were stained with combinations of murine monoclonal antibodies directly conjugated with fluorochromes (fluorescein isothiocyanate; FITC, phycoerythrin; PE, peridinin-chlorophyll protein; Per-CP or allophococyanin; APC). Staining was performed with antibodies specific for HLA class I antigens (HLA-A2 or HLA-A24) selectively expressed on INTUVAX cells, HLA-DR and CD86. The cells were incubated at 4 °C for 15 min with antibodies in the concentrations recommended by the manufacturer. The cells were analyzed in a FACSCanto™ II flow cytometer using the FACSDiva software and dot plots and quadrant statistics from a four-colour analysis were generated. All antibodies, the flow cytometer and the software were from BD Biosciences (Mountain View, CA, USA). Results for each subpopulation were expressed as the percentage of leucocytes. The samples were analyzed at the Dept of Clinical Immunology, Sahlgrenska University Hospital, Gothenburg, Sweden.

Serologic inflammatory parameters

Cytokines and chemokines were quantified from serum samples taken just before and 1 day after each administration of INTUVAX cells and 14 days after the second administration using a Bio-Plex human cytokine assay, a Bio-Plex 100 assay reader and the Bio-Plex Manager™ software (Bio-Rad Laboratories AB, Sundbyberg, Sweden). The analysis was performed according to the manufacturer’s instructions. The following inflammatory parameters were analyzed: IL-1 beta, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12p70, IL-13, IL-17A, G-CSF. GM-CSF, IFN-gamma, MCP-1, MIP-1 beta and TNF-alpha. The chemokine RANTES was quantified by a commercially available ELISA (R&D Systems, Abingdon, UK), according to the manufacturer’s instructions. The production of TNF-alpha, IL-1 beta, IL-12p70, RANTES and MIP-1 beta was also analysed in supernatants from thawed and washed Intuvax cells cultured for 24 h in AIM-V culture medium at a concentration of 1 × 10⁶ cells/mL. The samples were analyzed at the Dept of Clinical Immunology, Sahlgrenska University Hospital, Gothenburg, Sweden.

Lymphocyte distribution in peripheral blood

In order to evaluate potential changes in lymphocyte subset occurrence and activation state in peripheral blood taken just before and 1 day after each administration of INTUVAX cells and 14 days after the second administration, the following parameters were analyzed by FACS, as described above. Antibodies to the following antigens were used: CD3, CD4, CD8, CD19, CD16, CD56 and HLA-DR. The absolute number of blood lymphocytes was determined with Trucount reference beads using the method recommended by the manufacturer. The following subpopulations were reported CD3+, CD3 + 4+ and CD3 + 8+ T cells, CD3 + HLA-DR+ activated T cells, CD19+ B cells, CD3–16 + 56+ NK cells and CD3 + 16 + 56+ NKT cells. The results for each subpopulation were expressed as the percentage of lymphocytes and as the number of cells × 10⁹/L. All antibodies and the Trucount beads were from BD Biosciences.

Immunohistochemistry

A randomly selected tissue specimen from the renal tumor was formalin-fixed, paraffin-embedded and cut into 5 μm sections. The sections were deparaffinised, and rehydrated in a graded series of ethanols. Antigen retrieval was done by heating tissue sections, using a Target Retrieval Solution, pH 9.0 (DAKO). The samples were subsequently incubated for 5 min in peroxidase blocking solution (DAKO) to inhibit endogenous peroxidase. Consecutive sections from each tumor specimen were then incubated with antibodies against CD3 (polyclonal rabbit anti-human CD3; DAKO), CD4 (monolonal mouse anti-human CD4, clone 4B12; DAKO), CD8 (polyclonal rabbit anti-human CD8; DAKO), CD56 (monoclonal mouse anti-human CD56, clone 123C3; DAKO), CD68 (monoclonal mouse anti-human CD68, clone KP1; DAKO), HLA-DR (monoclonal mouse anti-human HLA-DR, alpha chain, Clone TAL.1B5; DAKO) and PD-L1 (monoclonal mouse anti-human PD-L1, Clone 22C3, DAKO-kit SK006, DAKO) for 30 min. A subsequent reaction was carried out using secondary antibodies (DAKO) at 37 °C for 30 min. Then, the sections were washed three times with phosphate-buffered saline and subsequently the color was displayed with DAB (DAKO) for about 5 min. Sections were counterstained with haematoxylin, dehydrated in ethanol, cleared in xylene, and coverslipped.

For enumeration of CD8-infiltrating T cells, five different areas with most abundant positively stained cells in each section were selected and number of positve cells were counted under 400 magnification [16]. The median number in each sample was then calculated. A semi-quantitative evaluation of infiltrating CD3+ T cells, CD4+ T cells, CD56+ NK/NKT cells, CD68+ macrophages and HLA-DR expressing non-tumor cells was performed. Moreover, the expression of HLA-DR by tumor cells was assessed.

ELISpot

The potential tumor-specific T cell response, induced by INTUVAX treatment was evaluated by running a tumor-specific IFN gamma (IFN-γ) ELISpot. Blood samples were collected at day 1 (before first vaccination) and 2 weeks after the second vaccination (in connection with hospitalization for nephrectomy). PBMC were isolated and frozen for subsequent batched analysis.

Fresh autologous tumor material from the resected tumor was deep-frozen (−70 °C) and subsequently thawed, homogenized with a GenltleMACS tissue homogenizer (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany), followed by six cycles of freezing at −130 °C and thawing. After centrifugation and sterile filtration of the supernatant, the method of Bradford was used to determine protein concentration. The lysate was used as an antigenic source in the tumor-specific ELISpot assay.

Freshly drawn PBMCs from each patient were used for differentiation of immature monocyte-derived DCs, as described above for the production of INTUVAX. Immature DCs were subsequently pulsed with the tumor material (100 μg/mL) described above and stimulated for 24 h with a cocktail identical with the one used for INTUVAX production.

A direct ex vivo interferon gamma ELISpot analysis was then performed, using the commercial Human IFN-γ ELISpot^PLUS kit (MABTECH AB, Nacka Strand, Sweden), according to instructions from the manufacturer.

Briefly, one vial of frozen PBMCs from each time point was thawed, counted in a cell counter (Sysmex K-4500; TOA Medical Electronics Co, Japan) and plated into the wells of an anti-IFN-γ antibody (clone 1-D1K, Mabtech, Inc) pre-coated ELISpot plate at 100,000 and 200,000 cells per well. Cells were then stimulated with 20,000 tumor-loaded autologous mature DCs per well for 24 h at 37 °C. As a positive control, an antibody to CD3 was used. After 24 h, plates were washed and incubated with a biotinylated anti-IFN-γ secondary antibody (clone 7-B6–1, Mabtech, Inc) for 2 h at room temperature. The plates were washed and incubated with streptavidin-conjugated alkaline phosphatase for 1 h and then washed again, and incubated with substrate for alkaline phosphatase. Excess substrate was removed by rinsing with tap water. The number of spots were counted using a dissection microscope. All samples were analyzed in triplicates, and the mean response (spots per 100,000 cells per well) was calculated. The analyses were performed at the Department of Clinical Immunology, Sahlgrenska University Hospital, Gothenburg, Sweden.

Measurement of potential auto- and alloimmunizaton

To evaluate potential autoimmune events, screening for autoantibodies against clinically relevant autoantigens, including nuclear antigens (ANA, SSA, SSB, Sm, RNP, Scl-70, Centromeres and Jo-1) and kidney parenchyma-associated autoantigens (liver-kidney microsomal antigens and mitochondrial antigens) was made. Samples were taken at screening and at day 120. The samples were analyzed at the Dept. of Clinical Immunology, Uppsala University Hospital, Uppsala, Sweden.

To evaluate potential vaccine-induced alloimmunization at the humoral level, screening for alloantibodies against HLA-A, B, C (MHC-class I) and HLA-DR, DQ, DP (MHC-class II) antigens was made. Samples were taken at screening and at day 120. The samples were analyzed at the Dept of Clinical Immunology at Uppsala University Hospital.

Tumor response

Tumor response was evaluated by CT scan, comparing the baseline CT scan (obtained within 28 days before day 1) with the CT scan made 3 months’ post nephrectomy, according to RECIST 1.1 criteria [17].

Patients

One patients (patient 6) with RCC and presumed RCC bone metastasis was later found to have multiple myeloma. This was not suspected until the patient developed progressive detoriation of the renal function, due to light chain damage. Retrospective analysis of the plasma proteins showed that myeloma was present before enrollment in this study. All 12 patients were included in both the efficacy and safety sets. However, patient 6 was not included in evaluation of tumor response or survival as the patient had 2 concomitant cancer diseases and thus not RCC with metastases.

INTUVAX characteristics and patient exposure

The manufactured INTUVAX batch passed quality (release) tests according to GMP guidelines, including sterility and endotoxin level (< 5 EU/mL). Number, viability, and HLA-DR expression was evaluated directly after thawing and the total number of viable and HLA-DR expressing cells/vial was found to be 12,6 million cells. The production of TNF-alpha, IL-1 beta, IL-12p70, MIP-1 beta and RANTES measured 24 h after thawing was 300, 800, 7.870, 6.460 and 29.000 pg/mL, respectively. When acceptance criteria (data not shown) of thawed INTUVAX cells were met, the rest of the frozen vials in the actual batch were transported from the CCK GMP laboratory to Vecura Clinical Research Center, Karolinska University Hospital, Stockholm, and transferred to a − 150 °C freezer for cell banking until time for vaccination.

On the day of administration/the day before (maximum 24 h before administration), the frozen vial was sent on dry ice to the local hospital pharmacy, where the final preparation of the INTUVAX product was made. Immediately before administration to the patient, the cells were thawed, washed and resuspended into final concentration of 10 or 20 × 10⁶ cells/mL in 0.15 M saline (Sodium Chloride; Braun Medical Inc.) with 2% human serum albumin (Albunorm; Octapharma). The estimated residual amount of R848, poly-I:C and IFN-gamma in each dose of the resuspended drug product was 0.25 ng, 2 ng and 0.1 Units, respectively. The administration of INTUVAX was performed within 1 h from thawing. All patients received 2 intratumoral administrations as per protocol. Patients 1–4 received 5 × 10⁶ cells in 0.5 mL/dose, patients 5–9 10 × 10⁶ cells in 0.5 mL/dose and patient 10–12 20 × 10⁶ cells in 1.0 mL/dose.

Safety

The patients were also extensively monitored in terms of laboratory assessments, and abnormalities in haematology, biochemistry, coagulation, and renal function, were in most cases assessed as minor without any signs of INTUVAX inducing clinically significant changes. These minor abnormalities and also the more extensive changes (e.g. albumin and calcium levels at CTCAE 3–4 for patient 8 and 11, respectively) were seen as a result of their current disease.

Tracking of injected INTUVAX cells

FACS analysis for donor (INTUVAX) cell tracking was performed after lysis of the erythrocytes on peripheral blood samples (taken just before vaccination and 1 h and 24 h after each vaccination), by staining with antibodies specific for the HLA class I antigens HLA-A2 or HLA-A24 that were selectively expressed on INTUVAX cells. The detection level for cell tracing was 0.0001% out of total circulating mononuclear cells (monocytes + lymphocytes) in blood. No tracking could be performed on patient 11 and 12, because the HLA-A tissue type of these two patients was A2 and A24 (thus identical to INTUVAX cells as to HLA-A tissue type). A measurable increase, corresponding to 8–35% of injected cells, was seen in the circulation 1 h and/or 24 h after administration in 4 out of 10 evaluated patients (patient 3, 8, 9 and 10).

HLA-compatibility between donor and recipient

Auto- and allo-immunization

Lymphocyte distribution and absolute number in peripheral blood

No significant changes in distribution or absolute number of CD3, CD4 or CD8 positive T cells, CD19+ B cells, CD3–16 + 56+ NK cells, CD3 + 16 + 56+ NKT cells or activated T cells (CD3 + HLA-DR+), NK cells (CD3–16 + 56 + 69+) or NKT cells (CD3 + 16 + 56 + 69+) were found 1 or 14 days after each administration of INTUVAX compared to base-line values before the first administration (data not shown).

Serological immune parameters

Inflammatory immune parameters in peripheral blood, taken just before INTUVAX administration and 24 h after each administration, were analysed in order to evaluate potential systemic release or induction of relevant cytokines, chemokines and other inflammatory mediators after intratumoral administration of INTUVAX. However, none of the evaluated inflammatory parameters in peripheral blood, including CRP, chemokines and cytokines exhibited any significant increase or decrease after vaccination (data not shown).

Tumor immunohistology

Examined tumor tissues from patient 2, 4, 8, 11 and 12 exhibited a massive infiltration (median > 200 cells/high power field, × 400 magnification) of CD8+ T cells (Fig. 1 and Table 5). The tumor tissue from patient 6 and 9 exhibited a strong infiltration (median 50–199 cells/high power field, repectively, while the tumor tissue from the other 5 patients exhibited a moderate (median < 20–49 cells/high power field) or weak (median < 20 cells/high power field) infiltration (Fig. 1 and Table 5). The infiltration of CD8+ T cells in 3 randomly selected tumors from RCC patients that were not involved in the study, was also assessed and all three samples exhibited a low to moderate number (< 50 cells/high power field) of infiltrating CD8+ T cells (data not shown).

Patient	CD8/HPF	Grade of CD8 infiltration
1	44	Moderate
2	>200	Massive
3	33	Moderate
4	>200	Massive
5	<5	Weak
6	73	Strong
7	47	Moderate
8	>200	Massive
9	132	Strong
10	30	Moderate
11	>200	Massive
12	>200	Massive

CD8/HPF, median number of intratumoral CD8+ T cells/high power field (×400)

As illustrated in Fig. 2, the number and infiltration pattern of CD8+ T cells were nearly identical to the number and infiltration pattern of CD3+ T cells within tumor cell nests and by far outnumbered CD4+ T in tumor samples with high intratumoral T cells infiltration. In contrast, CD4+ T cells generally predominated in the stroma surrounding tumor cell nests (Fig. 2b). The normal kidney tissue areas adjacent to tumor areas, including those adjacent to tumor areas with high CD8+ T cell infiltration, were generally only weakly infiltrated with CD8+ T cells (Fig. 2E). In the only evaluable metastatic lesion surgically removed (subcutaneous lesion removed 1 month after nephrectomy) from a patient with massive CD8+ T cell infiltration in the primary kidney tumor, a similar massive infiltration of CD8+ T cells was seen (Fig. 2F).

Only a few scattered CD56+ NK cells were found in the tumor tissue from all 12 patients. CD68+ macrophage-like cells were usually seen scattered between the tumor cell nests (data not shown). All tumor samples, except the tumor from patient 5 (with a papillary RCC), exhibited at least some areas of HLA-DR-expressing tumor cells. A general and very strong HLA-DR expression in tumor cells was found in tumors from patient 2, 8 and 12, all three with a concomitant massive infiltration of CD8+ T cells (Figs. 2 and 3). Staining for PD-L1 expression revealed that only one out of six evaluated tumors (patient 2, 4, 8, 10 and 12) exhibited a positive staining that exceeded 5% of all tumor cells. Notably, this tumor (from patient 10) was characterized by a low CD8-infiltration and weak or absent HLA-DR expression on tumor cells (Fig. 3). In contrast, none of the analysed tumors with massive CD8 infiltration and with a concomitant strong HLA-DR expression (patient 2, 4, 8,11 and 12) showed no or less than 5% of PD-L1 expression in tumor cells (Fig. 3).

Tumor-specific ELISpot

Nine out of 11 evaluated patients exhibited an increased number of tumor-specific and IFN-γ producing lymphocytes in peripheral blood, as determined by ELISpot, when comparing baseline values with values obtained 2 weeks after the second administration of INTUVAX (Fig. 4a). However, in ELISpot samples from 8 out of 11 evaluated patients (six with increased numbers post-administration and two with decreased numbers post-administration), the negative control sample (without addition of autologous tumor cell lysate) exhibited a number of spots that a least at one time-point exceeded the numbers found when the autologous tumor cell lysate was added. Taken together, formally reliable ELISpot results was only present in three patients, and these three exhibited an increased number of tumor-specific and IFN-γ producing lymphocytes 2 weeks after the second INTUVAX administration as compared to base-line levels (Fig. 4b).

Efficacy

Tumor response, as measured by comparing CT scan 3 months after nephrectomy with baseline by investigator assessment, according to RECIST 1.1 was evaluated in 8 patients. Patient 5 (incomplete metastasectomy during nephrectomy), patient 6 (myeloma lesions), patient 10 (deceased before the 3 months evaluation), patient 11 (not performed) were not evaluable. None of the 8 evaluable patients exhibited an objective tumor response, while 3 patients (patient 1, 2, 3) had stable disease and 5 patients (patient 4, 7, 8, 9, 12) had progressive disease.

Follow-up with standard of care

The study was formally completed in March 2014, but the patients have thereafter been followed-up for tumor progression, subsequent systemic treatment and tumor response to these treatment regimes and survival time.

Subsequent treatment with sunitinib

Two of 3 patients, who so far have received subsequent therapy with sunitinib as first-line TKI treatment, have shown an objective tumor response (Fig. 5), and both responding patients exhibited a strong/massive infiltration of CD8+ T cells in their removed kidney tumor. Both patients had a MSKCC poor prognosis profile at inclusion and both received two doses of 10 million vaccine cells. One of these 2 patients (patient 8) exhibited sarcomatoid histological features in the resected primary tumor and the other patient (patient 9) had developed 4 brain and 3 liver metastases at the follow up visit 3 months after nephrectomy. Notably, the patient with brain and liver metastases responded with a complete disappearance of all brain and liver lesions (Fig. 5) and this response was still ongoing 38 months after initiation of sunitinib treatment.

The third patient (patient 12), who received subsequent treatment with sunitinib, exhibited a strong/massive intratumoral infiltration of CD8+ T cells experienced a stable disease for 6 months on sunitinib, but progressed upon discontinuation (due to unacceptable side effects) of sunitinib treatment, despite initiation of axitinib treatment. This patient had an MSKCC poor prognosis profile at inclusion and received 2 doses of 20 million vaccine cells.

Subsequent treatment with immune checkpoint inhibitors

One patient (patient 1) has recently received immune checkpoint inhibitor (ICI) treatment. The patient started with nivolumab treatment 57 months after start of INTUVAX (28 months after start of subsequent TKI treatment) (Fig. 6). Follow-up CT scan 2 months after nivolumab start showed stable disease.

Overall survival data

Updated survival data (06 December 2016) on the whole patient group (poor + intermediate risk) showed that five of the eleven patients evaluable for efficacy were still alive. Median OS for the entire patient group was 42.5 + (yet not reached) months. For the MSKCC poor risk sub-group (6 patients), the median OS was 32.8 + months and for the subgroup of patients with MSKCC intermediate risk (5 patients), the median overall survival was 47.8 + months.