NovoTTF: where to go from here?

Glioblastoma (GBM) remains the most common and aggressive primary brain tumor in adults, and new therapies are desperately needed. Results using a device (NovoTTF, now Optune from Novocure) worn on the head that creates alternating electric fields (tumor treatment fields [TTFs]) in the brain to disrupt mitosis, which was tested in an internationally conducted phase III trial, were presented orally at the 2010 American Society of Clinical Oncology (ASCO) annual meeting and published in manuscript form 2 years later.¹ In that trial, 237 patients with recurrent GBM were randomized either to “best standard chemotherapy” (BSC) at the physician’s discretion (n = 120) or to the device (n = 117). The investigators and sponsor interpreted the results in a favorable manner, concluding that NovoTTF was non-inferior to various chemotherapies with less toxicity.¹ Various statistical methods were applied in analyzing the data, and controversy ensued, both at the ASCO meeting itself and afterward. Many, including us (T.F.C., A.B.L.), cautioned that it was also possible to interpret the results as demonstrating that: (i) neither TTF nor BSC were particularly effective for recurrent GBM, as the median progression-free survival (about 2 months in both arms) and the 6-month progression-free survival rate (21%, 95% CI: 13.5–29.3 for TTF vs 15%, 95% CI: 7.8–22.3 for BSC, P = .13) were poor and hardly exciting for either treatment; and (ii) failing to improve survival (the primary endpoint of the trial, about 7 mo in both arms, hazard ratio for death with TTF 0.86, 95% CI: 0.66–1.12, P = .27) is not statistically equivalent to affirming non-inferiority.¹ In our (T.F.C., A.B.L.) experience, nothing much changed in the field, despite FDA approval in 2011. Many interpreted the device as somewhere between a gimmick and a nontoxic placebo. (Of note, while the medical side effects were minimal and quality of life in fact was reported as favorable compared with chemotherapy,¹ the financial toxicity is not inconsequential.²) Personalized approaches targeting driver mutations continued as investigational drug trials, perhaps now renamed Precision Medicine, and immunotherapy trials blossomed.

In the interim, however, a randomized trial (EF-14) of NovoTTF was conducted in newly diagnosed GBM. Interim favorable results were presented orally at the 2014 Society for Neuro-Oncology (SNO) and 2015 ASCO annual meetings and published formally in JAMA.³ These interim results (among 315 randomized patients), like those of the trial for recurrent disease, were met with controversy and skepticism, even “polarizing.”⁴ Again, however, nothing dramatic changed in the field: other trials continued, few patients chose to wear the device, and concurrent or planned NovoTTF was exclusionary for trials testing novel agents.

Then, results of the “full data set” of 695 randomized patients (466 TTF, 229 control) were presented orally at the 2016 SNO annual meeting.⁵ The results show a numerical advantage in median overall survival with statistical significance and are hard to dismiss (Table 1). Certainly, we found them more compelling. In addition, NovoTTF became part of the National Comprehensive Cancer Network (NCCN) guidelines in the treatment of newly diagnosed GBM: “based upon lower-level evidence, there is uniform NCCN consensus that the intervention is appropriate” (Category 2A).⁶ The trial randomized adults with KPS ≥70 to radiotherapy and temozolomide followed by either adjuvant temozolomide alone or with NovoTTF. There was no sham device. The authors note, correctly, that early trials of radiotherapy in GBM similarly did not incorporate a sham device,⁷ yet the favorable results led to radiotherapy becoming accepted as standard of care. However, in our view, those trials were conducted in a different era, and a sham device could and should have been devised and implemented. Nonetheless, key results included median overall survival of 24.5 months with TTF versus 19.8 months without it, and a favorable hazard ratio of 0.65, a clinically meaningful and statistically significant improvement (Table 1).⁵ The survival advantage continued at later landmark timepoints such as the 2-year survival rate (from randomization) of 42.5% versus 30% (P = .001).⁵ We (T.F.C., A.B.L.) both offer the device as treatment, now with more enthusiasm than before these most recent results were publicized. But how do we put into perspective these results for clinical use, and incorporation toward future development of therapies for GBM?

Table 1.

Relative improvements in survival from clinical trials for newly diagnosed GBM

Study	Treatment	Number of Patients with GBM	Results								Absolute Improvement								Relative Improvement (%)								Fold Improvement
			mOS (mo)	mOS mMGMT (mo)	mOS uMGMT (mo)	1 year OS rate (%)	2 year OS rate (%)	3 year OS rate (%)	4 year OS rate (%)	5 year OS rate (%)	mOS (mo)	mOS mMGMT (mo)	mOS uMGMT (mo)	1 year OS rate (%)	2 year OS rate (%)	3 year OS rate (%)	4 year OS rate (%)	5 year OS rate (%)	mOS(mo)	mOS, mMGMT	mOS, uMGMT	1 year OS rate (%)	2 year OS rate (%)	3 year OS rate (%)	4 year OS rate (%)	5 year OS rate (%)	mOS(mo)	mOS, mMGMT	mOS, uMGMT	1 year OS rate (%)	2 year OS rate (%)	3 year OS rate (%)	4 year OS rate (%)	5 year OS rate (%)
Walker et al 1978*	Supportive Care	28	3.2	N/A	N/A	3	0	N/A	N/A	N/A	5.1	N/A	N/A	21	1	N/A	N/A	N/A		157%	N/A	N/A	700%	∞	N/A	N/A	N/A	2.6	N/A	N/A	8.0	∞	N/A	N/A	N/A
Walker et al 1978*	RT	63	8.3	N/A	N/A	24	1	N/A	N/A	N/A
Westphal et al 2006*	RT	106	11.4	N/A	N/A	45	7	0	N/A	N/A	1.7	N/A	N/A	15	5	2	N/A	N/A		15%	N/A	N/A	33%	71%	∞	NA	NA	1.1	N/A	N/A	1.3	1.7	∞	N/A	N/A
Westphal et al 2006*,****	Carmustine wafers & RT	101	13.1	N/A	N/A	60	12	2	N/A	N/A
Stupp et al 2009**	RT	286	12.1	15.3	11.8	50.6	10.9	4.4	3	1.9	2.5	8.1	0.8	10.5	16.3	11.6	9.1	7.9		21%	53%	7%	21%	150%	264%	303%	416%	1.2	1.5	1.1	1.2	2.5	3.6	4.0	5.2
Stupp et al 2009**	RT & TMZ	287	14.6	23.4	12.6	61.1	27.2	16	12.1	9.8
Stupp et al 2016***	RT & TMZ	216	19.8	21.2	13.9	65.3	30	15.9	10.4	N/A	4.7	8.5	3.4	7.3	12.5	7.6	6.9	N/A		24%	40%	24%	11%	42%	48%	66%	N/A	1.2	1.4	1.2	1.1	1.4	1.5	1.7	N/A
Stupp et al 2016***	RT & TMZ & TTF	456	24.5	29.7	17.3	72.6	42.5	23.5	17.3	N/A

Legend: GBM, glioblastoma; RT, radiotherapy; TMZ, temozolomide; TTF, Tumor Treatment Fields (NovoTTF device); mOS, median overall survival; N/A, not available; MGMT, O⁶-methylguanine-DNA methyltransferase; mMGMT, methylated MGMT promoter; uMGMT, unmethylated MGMT promoter; mo, months

*Each arm also intentionally accrued about 10% more patients with non-GBM malignant gliomas, not included here.

**Each arm included about 10% of patients who did not have GBM on central pathology review but were included in the intent-to-treat analyses and results.

***Median OS is from diagnosis; landmark survival is from randomization which occurred a median of 3.8 months after diagnosis.

****OS improvement with carmustine wafers was not statistically significant (Hazard ratio 0.78, p~0.1) in the GBM subgroup; landmark survival rates are estimated from Kaplan–Meier survival curve in the publication.

First, we appreciate the effort made in dogged pursuit of the trials, despite skepticism by colleagues, often healthy but occasionally bordering on derision and condescension. For a deadly and typically incurable disease like GBM, we would never want to discourage the development of new or alternative therapeutic approaches. Yet we often remain puzzled by the mechanism of action which, despite repeated explanation, does not lend itself to an easy understanding of the approach.

Second, no matter how compelling the results, and no matter how emphatically it is recommended, many if not most patients refuse NovoTTF in our (T.F.C., A.B.L.) clinical practice. The barriers we (T.F.C., A.B.L.) observe are several. Most importantly, the impact is incremental, not curative. The absolute size of clinical benefit associated with the addition of NovoTTF (~13% absolute improvement in 2-year overall survival rate from 30% to 43%)⁵ is not dissimilar to what was seen with the addition of temozolomide to radiotherapy (~16% absolute improvement in 2-year overall survival rate from 11% to 27%).⁸ However, the magnitude of effect from adding Novo TTF seems less profound when compared with studies resulting in the addition of radiotherapy⁷ or temozolomide⁸ to the standards of care in their era, especially when evaluating the later landmark survival timepoints (Table 1). For example, radiotherapy dramatically increased the 1-year survival rate 8-fold (Table 1). Perhaps more to the point, the 16% absolute improvement in the 2-year overall survival rate achieved by adding temozolomide to radiotherapy represented a 150% (and 2.5-fold) relative increase vs radiothearpy alone, whereas the 13% additional absolute improvement with NovoTTF and temozolomide and radiotherapy represents only a 42% (and 1.4-fold) relative increase vs temozolomide and radiotherapy (Table 1). Of note, NCCN guidelines include treatment with temozolomide, as “based upon high-level evidence, there is uniform NCCN consensus that the intervention is appropriate” (Category 1), a higher level than that assigned for NovoTTF (2A), although also based on a similar randomized phase III trial, reflecting the uniform acceptance of temozolomide vs remaining controversy surrounding NovoTTF.

Third, the subgroup of patients most likely to benefit is unclear. Our understanding of how O⁶-methylguanine-DNA methyltransferase (MGMT) promoter methylation impacts the benefit from temozolomide may be imperfect, yet the empiric observation remains that patients with tumors harboring methylated MGMT promoter benefit the most. In fact, the benefit of temozolomide is so limited among patients with GBMs that harbor “unmethylated” MGMT, particularly among the elderly, that many studies, some completed⁹ and others ongoing (eg, NCT NCT02383212, NCT02617589), now skip temozolomide altogether in patients with unmethylated tumors as a prudent way to speed development of therapies without unnecessary phase I studies in combination with temozolomide.⁹ As with other treatments, identifying the subgroup most likely to benefit from NovoTTF would provide a rational development plan for combination therapy and clarify for whom the effort associated with device compliance is most warranted.

Fourth, in our (T.F.C., A.B.L.) experience, some patients refuse to wear the device outright because it is a daily reminder, both private and public, of their illness. Some would far prefer to privately swallow a new pill at home, or receive a brief infusion every week or so, discreetly, without continuously announcing to their families, colleagues, and friends that they are dealing with brain cancer. In the JAMA discussion, this “burden” was identified as an explanation (inadequate in our view) for lack of a sham device, noting that “sham . . . treatment for the control group was considered neither practical . . . nor appropriate (due to the burden for patients and caregivers and the need to shave the scalp).”³ By contrast, other patients wear their device as a badge of honor, demonstrating an hourly effort in fighting their disease. In fact, one author (A.B.L.) had a Hassidic patient enthusiastically don the device over a shaved head, with payos glued to a yarmulke placed on top. Devices are in development that can be worn less obviously under clothing for thoracic and intra-abdominal systemic malignancies.

Fifth, an obvious issue is incorporation of NovoTTF into trials for newly diagnosed GBM. This should be considered. However, we also do not believe it is compulsory at this time. Carmustine eluting wafers¹⁰ (Gliadel) are another FDA-approved therapy listed in the NCCN guidelines⁶ that were popular for a period, but they are rarely used any longer, and omitting them from treatment is not considered unreasonable. We (T.F.C., A.B.L.) certainly would never recommend that a patient go back to the operating room for a second surgery exclusively to insert the wafers. Furthermore, trials for newly diagnosed GBM do not require wafer use, and nearly all trials exclude their use. From a safety perspective, mandating NovoTTF could necessitate new and lengthy phase I studies to address any interactions, no matter how unlikely, between the device and an experimental drug. Other companies pursuing combinations with NovoTTF might be additionally burdened with monitoring device compliance in attempting to achieve balance between trial arms, as duration of device use appears associated with efficacy. This would be a logistical challenge for a large phase III study and potentially confounding for smaller phase II studies (randomized or single arm). Additionally, requiring the device would impede accrual by dissuading many patients from participating in a trial of a new drug.

Finally, and perhaps most importantly, the EF-14 trial required patients to continue treatment beyond disease progression and was among the most intriguing aspects of the study design. It is interesting to speculate that we may see further incremental benefit from other therapies for taking this approach, and it makes us ponder whether we are too quick to abandon our interventions, especially in the absence of clearly effective alternatives. It also provides some insight into the effect of the therapy which might include changing the rate of progression (ie, slowed disease) in addition to (or instead of) tumor cell kill.

In short, the recently presented results of NovoTTF in newly diagnosed GBM⁵ are impossible to ignore. To the contrary, we should embrace all reasonable new therapies for GBM, including NovoTTF, which represents a treatment option for those patients interested in using it. NovoTTF use leads to an important but incremental survival improvement in GBM. However, further effort is needed toward optimizing the effect of this therapy. To us (T.F.C., A.B.L.), it is rational first to investigate variables related to physical settings of the device or array coverage as a way to better define predictors of benefit and mechanisms of escape. We would hope that the process of optimization would lead to more robust and durable efficacy, and to identification of a subgroup most likely to benefit. Given the incremental effect on survival, and other concerns outlined above, it is also our opinion that it is not necessary to incorporate NovoTTF into the clinical trial design for emerging therapies, much the way carmustine wafers are now rarely used and temozolomide can be deferred in clinical trials of patients with MGMT-unmethylated GBMs. However, NovoTTF is a new therapeutic modality that prolongs survival in newly diagnosed GBM, and as a field we should pursue efforts to optimize its use.

Funding

A.B.L. was supported in part by Cancer Center Support Grant P30 CA013696-42 and the NCI Community Oncology Research Program (NCORP) Project 1UG1CA189960-03. T.F.C. was supported in part by the Ivy Foundation, National Brain Tumor Society, Singleton Family Foundation, Uncle Kory Foundation, Ziering Family Foundation, and Art of the Brain.