The APOLLO investigators showed that next-generation sequencing of cerebrospinal fluid can reveal molecular alterations—how should this affect our management approach?

See related article by Xing et al., p. 6168

In this issue of Clinical Cancer Research, Xing and colleagues present findings from the APOLLO trial (1), a study that aims to gain a deeper understanding of the impact of osimertinib on the central nervous system (CNS). Targeted therapy using tyrosine kinase inhibitors (TKI) for the treatment of non–small cell lung cancer (NSCLC) harboring an activating mutation in the EGFR has completely changed the treatment approach for these patients. Despite remarkable response rates in the rest of the body, progression in the central nervous system (CNS) has remained a major problem for patients. The development of osimertinib, a TKI with preclinical evidence of CNS penetrance, was a major advance for patients. Since then, multiple trials have documented impressive CNS efficacy of this agent (2–4). In a subgroup analysis of the FLAURA trial, authors found that osimertinib 80 mg was associated with a superior CNS progression-free survival and CNS objective response rate (ORR) compared with erlotinib/gefitinib. In a retrospective composite of findings from patients with leptomeningeal disease treated on all of the AURA osimertinib development trials, investigators found that osimertinib 80 mg could lead to responses in leptomeningeal disease (ORR 55%). In the prospective BLOOM trial of patients with EGFR-mutated NSCLC harboring leptomeningeal disease, investigators found that osimertinib 160 mg led to an ORR of 62%. These studies, however, contained limited data on the in vivo CNS pharmacodynamic features of osimertinib, and the impact of these changes on effectiveness in the CNS.

The prospective APOLLO study enrolled 38 patients with NSCLC metastatic to the brain; exposure to an EGFR-tyrosine kinase inhibitor (TKI) and EGFR T790M detected on tissue testing were also required for enrollment. Patients were treated with osimertinib 80 mg and had plasma next-generation sequencing (NGS). Paired plasma and CSF NGS was completed on 12 patients at baseline, 6 weeks, and at the time of disease progression. The authors investigated the assumption that intracranial ORR is a result of osimertinib crossing the blood brain barrier by measuring the drug penetrance rate in the CSF. The concentration of osimertinib in the plasma and CSF at 6 weeks was strongly correlated, suggesting that higher plasma concentrations may lead to higher CSF concentration. The authors should be commended for prospectively following CSF analyses, and the patients should be applauded for their important contribution to our understanding of this complex disease. We believe that there are important clinical and translational conclusions that can be drawn from the APOLLO study.

From a clinical perspective, after the prior studies it was unclear if patients with leptomeningeal disease should receive 80 mg or 160 mg. Where then, does APOLLO leave our clinical question of dosing osimertinib in the case of leptomeningeal disease? When we consider the full population intracranial ORR across the relevant studies (55% and 68.8% for 80 mg daily and 62% for 160 mg daily), the immediate thought would be that dosing at the lower 80 mg is clearly adequate; response rates seem similar across the three studies regardless of dose. Yet when one considers the pharmacokinetic work done in the APOLLO study, it is important to remember that the investigators found that plasma concentration of osimertinib was clearly associated with CSF concentration. Given interpatient heterogeneity in drug metabolism, it may very well be true that some patients benefit from the higher dose of osimertinib. Indeed, recent retrospective data indicated that dose escalation of osimertinib in the setting of CNS progression was associated with an improvement in outcomes (albeit a modest one; ref. 5). Real-time pharmacokinetic measurement of plasma osimertinib concentration is beyond the scope of clinical practice. We would argue that while 80 mg is adequate for many patients, in the presence of further CNS progression a trial of dose escalation should be considered to determine if CSF concentration is the limiting factor.

From a translational perspective, the APOLLO study allows us to explore the potential role of CSF NGS to guide clinical decision-making in patients experiencing CNS progression. Any liquid biopsy, whether using plasma or urine, attempts to provide a more comprehensive view of the molecular profile of a patient's disease by sequencing tumor DNA from multiple sources. Yet this same effort to identify mechanisms of molecular resistance in multiple organs brings up a key limitation of liquid biopsies–sensitivity is lower than tissue assays (6). Does CSF function more like a tissue environment (high sensitivity for an alteration in that compartment) or like a liquid biopsy (lower sensitivity for alterations as leptomeningeal, parenchymal brain compartments are pooled)? In the APOLLO study, NGS on plasma samples yielded more alterations overall, but NGS on the CSF revealed additional copy-number alterations not seen in the plasma. The clinical relevance of copy-number alterations detected in liquid biopsies is controversial. However, the APOLLO authors did report one patient with EGFR T790M–positive tissue and CSF, but T790M negative on plasma assessment. Importantly, both patients with EGFR T790M detected in the CSF, including the patient with EGFR T790M–negative plasma, had a robust and durable intracranial response to osimertinib. This supports the concept that regardless of where T790M is detected it is a clear predictive biomarker of response to osimertinib (6).

The authors conclude that in cases where T790M is not detectable in plasma, CSF NGS should be considered. We are not sure we necessarily agree with that conclusion in this specific circumstance–prior meta-analyses have shown that T790M very rarely develops in CNS only progression (7). This finding is likely due to the lack of selective pressure to develop T790M from non-CNS penetrant first- and second-generation TKIs. Beyond this, every patient in the APOLLO study that was T790M positive in the CSF was also T790M positive in tissue, potentially undercutting the value of lumbar puncture in the setting of progressive disease accessible for tissue biopsy. Yet the principle underlying the authors' conclusion, that the CSF represents a unique compartment, can apply to other CNS penetrant TKIs used to target NSCLC with sensitive alterations such as ALK, ROS1, and RET. The literature on resistance alterations detectable in plasma and the development of drugs to target resistance alterations is growing quickly. Understanding how to detect these resistance alterations and whether the anatomic location of detection changes the clinical significance is critical to identifying the most appropriate treatment in these patients. We believe that CSF NGS may be much more useful in identifying mechanisms of resistance to CNS-penetrant TKIs, where selective pressure is more likely to induce a targetable resistance mechanism.

One of the great boons of evaluating plasma ctDNA is the ease of access to this information rich source. CSF does not share this advantage and patients and physicians may, understandably, resist sampling this space. Nevertheless, the APOLLO investigators clearly show that CSF sampling has clinical value in the case of CNS progression. We would thus favor considering CSF evaluation in the same vein as tissue biopsy–if a plasma ctDNA assay is inconclusive, NGS testing should be pursued in the organ system most directly involved in the progression. If the patient has growing liver metastases, that means a liver biopsy. If the patient has CNS-only progression, that means CSF sampling (Fig. 1).

Figure 1.

A proposed schematic for the use of plasma ctDNA, CSF, and tissue-based NGS testing for progressive NSCLC.

Figure 1.

A proposed schematic for the use of plasma ctDNA, CSF, and tissue-based NGS testing for progressive NSCLC.

Close modal

M.E. Marmarelis reports grants from Eli Lilly, Trizell, and AstraZeneca, personal fees from Novocure and Boehringer Ingelheim, nonfinancial support from Novartis (medical writing), and other from Gilead Sciences (stock), Portola Pharmaceuticals (stock), Merck (stock), Bluebird Biosciences (stock), Johnson & Johnson (stock), and Pfizer (stock) outside the submitted work. J.M. Bauml reports grants and personal fees from Merck, Clovis, Janssen, AstraZeneca, Takeda, and Incyte, grants from Carevive Systems, Novartis, and Bayer, and personal fees from Bristol-Myers Squibb, Celgene, Boehringer Ingelheim, Guardant Health, Genentech, Ayala, Regeneron, and Inivata outside the submitted work. No other potential conflicts of interest were disclosed.

J.M. Bauml has received grant support from the Lungevity Foundation.

1.
Xing
L
,
Pan
Y
,
Shi
Y
,
Shu
Y
,
Feng
J
,
Li
W
, et al
Biomarkers of osimertinib response in patients with refractory, EGFR-T790M-positive non-small cell lung cancer and central nervous system metastases: the APOLLO study
.
Clin Cancer Res
2020
;
26
:
6168
75
.
2.
Reungwetwattana
T
,
Nakagawa
K
,
Cho
BC
,
Cobo
M
,
Cho
EK
,
Bertolini
A
, et al
CNS response to osimertinib versus standard epidermal growth factor receptor tyrosine kinase inhibitors in patients with untreated EGFR-mutated advanced non–small-cell lung cancer
.
J Clin Oncol
2018
;
36
:
3290
7
.
3.
Ahn
M-J
,
Chiu
C-H
,
Cheng
Y
,
Han
J-Y
,
Goldberg
SB
,
Greystoke
A
, et al
Osimertinib for patients with leptomeningeal metastases associated with epidermal growth factor receptor T790M positive advanced NSCLC: the AURA LM analysis
.
J Thorac Oncol
2020
;
15
:
637
48
.
4.
Yang
JCH
,
Kim
S-W
,
Kim
D-W
,
Lee
J-S
,
Cho
BC
,
Ahn
JS
, et al
Osimertinib in patients with epidermal growth factor receptor mutation–positive non–small-cell lung cancer and leptomeningeal metastases: the BLOOM study
.
J Clin Oncol
2020
;
38
:
538
47
.
5.
Piper-Vallillo
A
,
Rotow
JK
,
Aredo
JV
,
Shaverdashvili
K
,
Luo
J
,
Carlisle
JW
, et al
High-dose osimertinib for CNS progression in EGFR+ non-small cell lung cancer (NSCLC): a multi-institutional experience
.
J Clin Oncol
2020
;
38
:
9586
.
6.
Oxnard
GR
,
Thress
KS
,
Alden
RS
,
Lawrance
R
,
Paweletz
CP
,
Cantarini
M
, et al
Association between plasma genotyping and outcomes of treatment with osimertinib (AZD9291) in advanced non-small-cell lung cancer
.
J Clin Oncol
2016
;
34
:
3375
82
.
7.
Wang
ZF
,
Ren
SX
,
Li
W
,
Gao
GH
. 
Frequency of the acquired resistant mutation T790 M in non-small cell lung cancer patients with active exon 19Del and exon 21 L858R: a systematic review and meta-analysis
.
BMC Cancer
2018
;
18
:
148
.