Intralesional Therapies for Metastatic Melanoma

Sanjiv S. Agarwala, MD

Professor of Medicine
Temple University School of Medicine
Chief, Oncology and Hematology
St. Luke’s Hospital and Health Network
Bethlehem, Pennsylvania

Since my 2012 report in The Melanoma Letter on promising intralesional therapies for metastatic melanoma,1 researchers have achieved important milestones in the field. Foremost were the October 2015 U.S. Food and Drug Administration’s approval and January 2016 European approval of talimogene laherparepvec (T-VEC, brand name Imlygic,® formerly OncoVEXGM-CSF), an oncolytic immunotherapy derived from HSV (herpes simplex virus)-1. T-VEC received approval in the U.S. for melanoma patients with inoperable tumors, and in Europe for melanoma patients with certain inoperable tumors.

Researchers have also been accumulating significant clinical efficacy findings with other intralesional strategies and gathering evidence yielding insights into mechanisms behind their systemic effects. 


Not all the news is good. Our initial report in The Melanoma Letter noted that in a phase 2 trial, one of the experimental intralesional therapies, Allovectin-7 (velimogene aliplasmid), a plasmid/lipid complex, had produced an objective response rate of 12 percent and stable disease in 25 percent of patients, without grade 3 or higher toxicities. Its phase 3 trial versus the chemotherapy combination dacarbazine (DTIC)/temozolomide (TMZ) in 390 patients with recurrent stage III or IV melanoma was ongoing. In more recent phase 3 trial reporting, the overall response rate at/or >24 weeks, the primary endpoint, was only 4.6 percent for velimogene aliplasmid and 12.3 percent for DTIC/TMZ (P=.01). Duration of response among velimogene aliplasmid responders was marginally longer (P=.066), but overall survival, a median 18.8 months, was shorter [95% CI 16.6, 21.3] than the 24.1 months [17.1, 27.9] (P=.491) with DTIC/TMZ. The results led the authors to conclude that for the selected population, velimogene aliplasmid was not an effective treatment, and researchers discontinued its development program.

The Gains with T-VEC

We had also reported on the ongoing phase 3 OPTiM trial of T-VEC (designed to secrete the cytokine GM-CSF, granulocyte macrophage-colony stimulating factor) versus subcutaneous GM-CSF alone in stage IIIB/IV melanoma patients. OPTiM included 436 patients (median age 63 years, 57 percent male), with 295 receiving T-VEC and 141 receiving GM-CSF alone. Durable response rate (complete or partial response lasting continuously for at least six months by independent review) and overall survival were the primary endpoints.

The research moved ahead successfully after that. The primary OPTiM results, presented at the American Society of Clinical Oncology (ASCO) annual meetings in 2013 and 2014, and published in the Journal of Clinical Oncology in 2015, included a durable response rate (DRR) with T-VEC of 16.3 percent (48/205) (95% CI: 12.1%, 20.5%) vs only 2.1 percent with GM-CSF (3/141, P<.0001) (95% CI: 0%, 4.5%).2 The overall response rate (ORR) with T-VEC was 26.4 percent (95% CI: 21.4%, 31.5%), with 10.8 percent complete responses (CR). For GM-CSF the ORR was only 5.7 percent (95% CI: 2%, 10%) with <1 percent CR.

Median overall survival (OS) with T-VEC, 4.4 months longer than with GM-CSF alone, was 23.3 months (95% CI, 19.5 to 29.6 months), vs 18.9 months with GM-CSF alone (95% CI, 16.0 to 23.7 months). The finding achieved near-statistical significance (P=.051) with a hazard ratio of 0.79 (95% CI, 0.62 to 1.00).

An exploratory analysis of relative effects of the treatment yielded some noteworthy differences across patient subgroups. In patients with stage IIIB or IIIC melanoma, durable response rate differences between the T-VEC and GM-CSF arms were more pronounced (33 percent vs 0 percent), as they were with stage IVM1a disease (16 percent vs 2 percent). The differences were not as striking for patients with stage IVM1b (3 percent vs 4 percent) or stage IVM1c disease (7 percent vs 3 percent).

Among treatment-naïve patients, this pattern was also clearly evident, with DRRs of 24 percent for T-VEC and 0 percent for GM-CSF patients. However, the differences were less pronounced in the second-line setting and beyond: 10 percent with T-VEC vs 4 percent with GM-CSF. Similarly, the OS difference between T-VEC and GM-CSF was greater in stages IIIB, IIIC or IVM1a (HR, 0.57; 95% CI, 0.40 to 0.80). The authors speculated that these disparities may reflect the differences between locoregional and visceral disease: With earlier-stage, locoregional disease, both lytic and systemic immune effects are operant, but with visceral disease only systemic immune effects come into play, and to date, T-VEC has not proven to have significant systemic effects. Tumor reductions of ≥50 percent with T-VEC were observed in only 15 percent of uninjected measurable visceral lesions.

Fatigue (T-VEC 49 percent, GM-CSF 36 percent), chills (T-VEC 49 percent, GM-CSF 9 percent) and pyrexia (T-VEC 43 percent, GM-CSF 9 percent) were the most common adverse events (AEs). Discontinuations for AEs were uncommon, with 4 percent of T-VEC patients and 2 percent of GM-CSF patients discontinued. The only grade 3 or 4 AE occurring with T-VEC (in >2 percent of T-VEC–treated patients) was cellulitis (2.1 percent).

The authors emphasized that OPTiM, to their knowledge, “is the first randomized controlled phase 3 study evaluating an oncolytic immunotherapy to demonstrate a therapeutic benefit in melanoma.”

An OPTiM study extension3 of randomized treatment including 30 patients (27 receiving T-VEC; 3 receiving GM-CSF), reported at the 2014 European Society for Medical Oncology (ESMO) meeting, revealed continuing improvement among patients receiving T-VEC but not those on just GM-CSF. Best overall responses improved in seven patients in the T-VEC group, with five patients who had partial responses in the main trial achieving complete responses, and two patients with stable disease in the main trial also achieving complete responses. With ongoing T-VEC treatment, complete resolution was achieved in a quarter of the patients, at least stable disease in 86 percent and a new durable response in one patient.

Electroporation of Plasmid Interleukin-12 (IL-12)

In electroporation, scientists apply an electrical field to cells to increase their permeability so that certain agents or coding DNA can be introduced into them therapeutically. In melanoma treatment, physicians can use an electrode to open tumor cell pores to allow a higher influx of the cytokine interleukin-12 (IL-12) for a longer time span than would occur with simple administration of IL-12 as a systemic therapy. This administration route also helps by comparatively reducing systemic IL-12 concentrations necessary for therapy. The IL-12 promotes antitumor activity by augmenting adaptive and innate immune responses, among other mechanisms. Specifically, it enhances the immune capacity of NK (natural killer) cells and T cells, upregulating interferon (IFNγ) as well as antigen presenting and processing. Electrochemotherapy is widely available in Europe.

In a phase 1 dose-escalation safety study of intratumoral electroporation of plasmid IL-12, presented at the 2014 Melanoma Bridge Meeting by Daud, et al,4 median long-term OS among 24 melanoma patients was 24.2 months. Among the nine patients whose treatment led to stable disease (SD) or better, OS was 46.4 months. Systemic responses to the local plasmid IL-12 therapy, as evidenced by SD or objective regression of noninjected lesions, were observed in 53 percent of patients (10/19) with metastatic disease. Without any concurrent or subsequent systemic therapy, regression of all distant lesions was complete in 11 percent of patients (2/19). Grade 1 and 2 procedure-related transient pain, the most frequent adverse events, occurred in 54 percent and 46 percent of patients, respectively.

The study, launched in 2007, was the first to assess gene therapy delivered locally in humans via electroporation with the aim of using DNA plasmid to induce systemic antitumor immunity. Plasmid IL-12-induced disease stabilization correlated with improved survival, the authors said. The results suggest the potential for both monotherapy and combination applications.

A phase 2 study of IL-12 electroporation among 28 patients with advanced melanoma offered promising results. The schedule called for IL-12 injections on days 1, 5 and 8 for a maximum of four cycles at 12-week intervals, with a primary endpoint of best overall response rate within 24 weeks of first treatment. In those receiving at least one treatment cycle, the results (based on modified RECIST criteria) were 32.2 percent of patients with an objective response and 10.7 percent of patients with a complete response. The CR rate for 85 evaluated lesions was 44.7 percent, with partial responses (PRs) in 8.2 percent of lesions and SD in 30.6 percent. In 22 patients with evaluable untreated distant lesions, regression was reported in 13 (59.1 percent). There were no serious adverse events. Injection site pain (69 percent of patients) and inflammation (20.7 percent of patients) were all grade 1 and 2, except for one report of grade 3 pain. An expansion protocol is planned. 


In some cancer tumors (melanoma, non-small cell lung, bladder, breast and prostate tumors included), surface intercellular adhesion molecule-1 (ICAM-1) is upregulated. Coxsackievirus A21 (CVA21) is a naturally occurring “common cold” ICAM-1-targeted RNA virus. CVA21-lysed tumor cells have been shown to induce a secondary systemic host-generated antitumor immune response in animal models.

Rates of immune-related progression-free survival (irPFS) were favorable for CAVATAK in CALM (CAVATAK in Late-Stage Melanoma), a phase 2 trial in patients with stage IIIC and IV melanoma. Andtbacka, et al5 reported responses in both injected and uninjected lesions. CALM included 57 stage IIIC and stage IV patients (42.1 percent stage III, 57.9 percent stage IV), 36 of them male, all with at least one injectable dermal, cutaneous, subcutaneous or lymph node lesion. The proportion of patients at six months with a complete response, partial response or stable disease (irPFS) (by irRECIST 1.1 criteria) was the primary endpoint, and overall response rate (complete response + partial response) was the secondary endpoint.

While investigators had initially hoped that at least 10 patients would achieve irPFS, at final analysis the total was 22 of the 57 enrolled patients (38.6 percent). The overall response rate was 28.1 percent (16/57 [8 CR + 8 PR]). Other noteworthy endpoints included a durable response rate of 21.1 percent, median time to response of 3.4 months (95% CI: 1.5, 4.2), median overall survival of 26 months (95% CI: 16.7, NR) and one-year survival of 75.4 percent (43/57).

There were no grade 3 or grade 4 treatment-related adverse events. Overall, multidose intralesional therapy with CVA21 was well tolerated.

The investigators observed responses in injected lesions, uninjected nonvisceral lesions and distant uninjected visceral lesions. To learn if observed tumor responses in uninjected lesions were immune-related or a consequence of CVA21 virus entering the tumor and stimulating a response, CALM investigators determined that virtually all patients had developed neutralizing antibodies to CVA21 by around day 22 after the fifth CAVATAK injection. Initially, 85 percent or more did not have neutralizing antibodies. The objective responses, despite the presence of the antibodies, suggests that the responses were likely immune-related.

Andtbacka, et al6 concluded that use of CVA21 in combination treatment with CTLA-4 or PD-1 checkpoint inhibitors such as ipilimumab or pembrolizumab, or with targeted small molecules such as BRAF and MEK kinase inhibitors, might result in enhanced antitumor activity. Clinical evaluation of CVA21 administered both intralesionally and intravenously in combination with ipilimumab in patients with unresectable melanoma is ongoing.

PV-10 (Rose Bengal)

Our original overview of intralesional therapies detailed interim findings of a phase 2 80-patient trial of PV-10 in stage III and stage IV melanoma, showing a 24 percent complete response rate in both target and bystander lesions, with loco-regional disease control in 71 percent of target lesions and 55 percent of bystander lesions. Regression of bystander lesions strongly correlated with response in target lesions. The objective response rate (complete plus partial response) was 49 percent in target lesions and 33 percent in designated bystander lesions. Analysis of the first 40 patients with complete responses showed PFS to be longer (11.1 months) than in those with stable disease or progressive disease (2.8 and 2.7 months, respectively).

In subsequent reporting of the final results in 2012, the objective response rate had climbed to 58 percent in target lesions and to 40 percent in bystander lesions. Locoregional disease control had increased to 80 percent for target lesions and 60 percent for bystander lesions. The close correlation between response in injected lesions and response in untreated “bystanders” persisted. Also, stasis or regression in distant visceral lesions was noted in several subjects.

Our stratification of target lesion findings according to disease stage7 revealed consistently robust responses to PV-10 in stage III melanoma subjects. It showed also that response duration was significantly longer in stage III patients than in stage IV patients, a mean of 9.6 months in stage III compared to 3.1 months in stage IV (P<.001). Greater baseline tumor burden in stage IV patients adversely affected response, as did progression of non-study lesions that precluded repeat treatment.

Guided by these observations, we restricted inclusion into the now ongoing phase 3 trial of PV-10 to subjects with stage IIIB-IIIC disease. In our phase 2 subgroup analysis of the original 28 patients in whom all lesions were injected with PV-10 (ASCO 2014), we had seen enhanced benefits. The complete response rate was 50 percent (CI 31-70 percent) and the overall response rate was 71 percent (CI 51-87 percent). A group of 26 patients added to the phase 2 analysis, among whom all had one or two monitored, uninjected bystander lesions, had a complete response rate of 64 percent (232/363) in injected lesions and a 36 percent complete response rate (10 of 28) in uninjected bystander lesions.

While we were conducting our preliminary phase 3 clinical trials, we undertook additional investigations into how local ablation may be activating systemic responses. Since responses in untreated lesions have occurred only when responses have occurred in injected lesions, and since these bystander lesion responses are typically delayed relative to the responses in the injected lesions, it suggests an immune-mediated process. A hypothetical mechanism for systemic immune activation for all the intralesional therapies that induce tumor lysis is that tumor lysis exposes antigenic tumor fragments to antigen-presenting cells, leading to specific T cell responses. Insights from these studies may suggest means to further enhance responses.

A Pilon-Thomas, et al8 murine study sought to determine if the lysis of distant tumor cells was caused by systemic distribution of PV-10 or by induction of a T cell response that spread systemically. (See Figure 1.)

PV-10 Immuno-Chemoblation
Figure 1. PV-10 Immuno-Chemoblation

When the researchers treated induced tumors with PV-10 or placebo, mean tumor size shrank in the PV-10-treated tumors by two-thirds (to 100 mm2 from about 300 mm2, P<.001) and bystander lesion size by about 38 percent (to 220 mm2 from about 300 mm2, P<.05). Survival and production of IFN-γ, a cytokine critical for innate and adaptive immunity (including tumor control) and for activating macrophages, were significantly higher with PV-10 treatment.  

A murine adoptive transfer experiment by Pilon-Thomas, et al further demonstrated that T cells purified from the spleens of B16 tumor-bearing mice treated with intralesional PV-10 were activated against B16 melanoma tumors in other mice.8 Other studies showed that direct injection of PV-10 into tumors is necessary to produce a systemic effect. Pilon-Thomas concluded that the studies confirmed PV-10 chemoablation’s direct effects on injected lesions as well as systemic effects leading to regression of uninjected subcutaneous and distant lesions.

Moving from murine to human research, a pilot study by Sarnaik, et al (ASCO 2014)9 among a small group of human subjects (n=13) showed significant increases in circulating CD3+ and cytotoxic CD8+ T cells after PV-10 tumor ablation, suggesting an immune-mediated antitumor response.

Combination Therapy

During the developmental testing phases for intralesional melanoma therapies, researchers have commonly espoused the view that their ultimate use would be in combination with the emerging systemic immunotherapy agents (e.g., anti-CTLA-4, anti-PD-1, anti-PD-L1). It is indeed very plausible that by evoking the release of tumor-derived antigens, intralesional therapies could enhance the efficacy of the T cell responses triggered by the systemic immunotherapies. Keeping in mind the intralesional therapies’ low and non-overlapping side effects, such combination therapies might well prove more effective than either agent alone.

Preclinical Combination Research

Murine studies with a melanoma model of intralesional PV-10 combined with a systemic anti-CTLA-4 analog (9H10)10 showed significantly reduced lung metastases compared with the systemic 9H10 alone. Separately, flank tumor size increased with 9H10 alone, but became no longer measurable with the combination. A further study showed longer overall survival with the combination versus 9H10 alone.

Clinical Combination Research

The promising animal research has helped lead to human trials. In phase 1b testing of T-VEC combined with ipilimumab11 in 19 treatment-naïve stage IIIB-IV melanoma patients, Puzanov, et al showed an objective response rate of 50 percent with durable responses in 44 percent, complete responses in 22 percent and a 72 percent disease control rate. The 70-patient phase 2 portion of the trial is ongoing.

A phase 1b/2 clinical trial of PV-10 in combination with pembrolizumab, an approved anti-PD-1 therapy, is ongoing among patients with stage IV metastatic melanoma, with completion of the 1b portion expected this year. In the subsequent phase 2 study, participants will be randomized 1:1 to the combination of PV-10 and pembrolizumab or to pembrolizumab alone (i.e., PV-10 + standard of care vs. standard of care).

At the same time, a phase 3 international multicenter, open-label, randomized controlled trial of single-agent intralesional PV-10 monotherapy is ongoing with systemic chemotherapy as the comparator. Patients are BRAF V600 wild-type and have locally advanced cutaneous melanoma. Treatment with ipilimumab or another immune checkpoint inhibitor has failed in these patients or they are not candidates for such treatment. Subjects in the PV-10 arm receive intralesional PV-10 to all of their melanoma lesions, while patients in the comparator arm receive the investigator’s choice of dacarbazine or temozolomide. Progression-free survival is the primary outcome.


The recent approval of T-VEC as an intralesional therapy for advanced melanoma is a significant milestone. We need to keep in mind, however, that melanoma is a systemic disease. The role for intralesional therapies as monotherapies is therefore limited mostly to use among those who are either not candidates for systemic therapy or whose disease has advanced despite systemic therapy.

A circumstance that may amplify regional use of intralesional therapies is the higher incidence of acral lentiginous melanoma in locales where populations of Asian and African nonwhites are prevalent. There, local regional spread and accompanying morbidity are often the main problems facing clinicians, with both persisting for many years.

Since pembrolizumab and nivolumab as monotherapies were approved by the FDA in late 2014 for treatment of advanced melanoma, there have been half a dozen further approved extensions of indications for these CTLA-4 and PD-1 checkpoint inhibitors, including combinations of both, with approvals of PD-L1 inhibitors looming in the near future. Patients whose tumors harbor the BRAF mutation have even more options with the approved targeted therapies.

The near future of intralesional therapies will be defined by ongoing and planned combination trials adding direct injections of intralesional therapies to systemic immunotherapies via checkpoint inhibitors. Potentially, they may be paired with targeted therapies as well. Initial findings from early-phase clinical trials of intralesional therapies combined with systemic immunotherapies suggest that response rates to these combinations improve beyond those historically achieved with the individual components alone. Yet we have to wait a while longer to see if survival benefits accrue and if recurrence rates go down. If they do, that will support the notion that early use of intralesional therapies may strengthen systemic immune responses.


  1. Agarwala S. PV-10, aka rose bengal: intra-lesional therapy for metastatic melanoma. The Melanoma Letter 2012; 30:1:5-7.
  2. Andtbacka RHI, Collichio FA, Amatruda T, et al. OPTiM: A randomized phase III trial of talimogene laherparepvec (T-VEC) versus subcutaneous (SC) granulocyte-macrophage colony-stimulating factor (GM-CSF) for the treatment (tx) of unresected stage IIIB/C and IV melanoma. LBA9008, ASCO 2013. Journ Clin Oncol 2013 ASCO Annual Meeting Abstracts; 31:18 suppl (June 20 Supplement).
  3. Nemunaitis JJ, Antbacka RHI, Ross M, et al. Results of the extension trial of OPTiM, a multicenter, randomized phase 3 trial of talimogene laherparepvec (T-VEC) vs GM-CSF for unresected stage. . . Abstract 1102P. ESMO 2014. Madrid, Spain. Ann Oncol 2014; 25(suppl 4):iv374-iv393.  10.1093/annonc/mdu344.
  4. Daud A, Takamura AT, Diep Tu, Heller R, Pierce RH. Long-term overall survival from a phase I trial using intratumoral plasmid interleukin-12 with electroporation in patients with melanoma. J Transl Med 2015; 13(Suppl 1): O3.
  5. Andtbacka RHI. CALM study: a phase 2 study of the efficacy and safety of intratumoral CAVATAK (Coxsackievirus A21, CVA21) in patients with Stage IIIc and Stage IV malignant melanoma. Abstract 1103. ESMO 2014, Madrid, Spain.
  6. Andtbacka RHI, Curti BD, Kaufman H, et al. Final data from CALM: a phase II study of Coxsackievirus A21 (CVA21) oncolytic virus immunotherapy in patients with advanced melanoma. Abstract 9030, ASCO 2015. Chicago, IL. J Clin Oncol 2015; 33 (suppl; abstr 9030).
  7. Agarwala S. PV-10. Second European Post-Chicago Melanoma meeting, Munich, Germany, June 2012.
  8. Pilon-Thomas SA, Weber A, Kodumudi K, et al. Intralesional injection with PV-10 induces a systemic anti-tumor response in murine models of breast cancer and melanoma. Abstract #1248, AACR 2013, Washington, DC.
  9. Sarnaik A, Crago G, Liu H, et al. Assessment of immune and clinical efficacy after intralesional PV-10 in injected and uninjected metastatic melanoma lesions. Abstract 9028, ASCO 2014. Chicago, IL.
  10. Wachter EA, Blair SO, Singer JM, Dees HC. Combination of PV-10 immuno-chemoablation and systemic anti-CTLA-4 antibody therapy in murine models of melanoma. Abstract 4755, AACR 2013, Washington, DC.
  11. Puzanov I, Milhem MM, Minor D, et al. Talimogene laherparepvec in combination with ipilimumab in previously untreated, unresectable stage IIIB-IV melanoma. Downloaded from on July 27, 2016 from