Herein, we report the characteristics of seven patients with choroidal metastasis from a cohort of 83 consecutive EGFR-mutant NSCLC patients (8.4%). Based on published data, the prevalence of choroidal metastases in NSCLC patients has been reported to vary from 0.2%1,8 to 7%9 in historical series, reaching 10% in post-mortem histopathological studies.10 The specific prevalence of choroidal mutation in EGFR-mutant NSCLC patients is still unknown. We take charge of nearly 300 new metastatic lung cancer patients each year in our department, pertaining to a 945-bed tertiary University Hospital in the North of Paris, covering a population of roughly 4.0 million inhabitants. Thus, 900 Stage IV lung cancer patients were treated during the same time period when these seven cases were registered. Of note is that all choroidal metastasis patients but one exhibited a tumor EGFR-activating mutation, the remaining patient showing a tumor K-Ras mutation. Among the seven patients, six were symptomatic, which was concordant with Shields’ cohort of 420 patients with various primary cancers over a 20-year period,1 of whom most were symptomatic. The choroidal metastases were inaugural in these six symptomatic cases, while being a revealing symptom in the first patient described.
Our seven-patient series only differed in age from the non-choroidal metastases population, showing a younger mean age. Most choroidal metastases concerned women, which is not unexpected taking account for the larger female prevalence in patients with EFGR-mutant tumor. However, if lung cancer is the most common source of choroidal metastases in males (40%), it is the second most common etiology in females, after breast cancer (68%).1
Concomitant TP53 mutation and Rb expression loss were unlikely to be predictors of choroidal metastasis, although it has been previously associated to SCLC histological transformation, since only three out of seven patients showed TP53/Rb double alterations with only a single patient exhibiting SCLC transformation upon EGFR TKI in our series.
In the United States, the frequency of EGFR testing was reported to be only 22% for Stage IV lung adenocarcinoma in the whole country in 2010,5 though the guidelines recommend systematic testing for all Stage IV lung adenocarcinomas.11 Although the rate of EGFR testing may have improved meanwhile, the occurrence of choroidal metastases should strongly support EGFR testing, considering EGFR TKI efficacy. Our patients had at least one more metastatic site, which is consistent with Kreusel et al.’s series, in which the only risk factor of developing choroidal metastases in univariate analysis was the presence of at least two other metastatic sites.9
In our series, a systematic control ophthalmological examination upon EGFR-TKI treatment was not performed and we could not correlate symptom decrease and choroidal lesion disappearance on dilated fundus ophthalmoscopy. Conversely, sub-retinal detachment, on echography or fluorescein angiography, frequently disappeared.
A retrospective analysis of brain MRI by an expert radiologist, who was blinded to the choroidal metastasis diagnosis, identified choroidal metastases in five patients, with a previously known diagnosis in four, and a previously unknown in the remaining one. The choroidal metastases’ appearance upon MRI consisted of a well-demarcated mass iso-intense on T1-weighted images and hypointense on T2-weighted images.12 Indeed, MRI has proven useful in differentiating choroidal metastases from primary uveal neoplasm, especially when ophthalmological visualization is obscured by subretinal effusion.13,14 Since brain imaging is essential to the diagnostic assessment of Stage IV NSCLC, a specific examination focused on the eyes should be recommended, particularly when EGFR-mutant lung cancer patients experience visual symptoms, such as visual disturbance.
Among four out of five evaluable patients treated by EFGR TKI, this treatment yielded a dramatic response with diminution and even disappearance of visual troubles after a 1-month course. Conversely, one patient experienced regression of visual troubles upon chemotherapy prior to TKI initiation. Tumor progression secondary to TKI resistance was not associated with choroidal progression, and choroidal response to TKIs appeared to be maintained, even in case of extra-choroidal progression.
The first described choroidal metastasis cases in possibly EGFR-mutant Asian patients (not molecularly assessed), published in 2009–2010, revealed the efficacy of EGFR TKI (in association with intravitreal bevacizumab).15,16 Since then, a few case reports have confirmed the ability of first- or second-generation TKIs to induce choroidal metastasis regression in patients with tumor EGFR mutation.7,17,18 The efficacy of third-generation EGFR TKI osimertinib on choroidal metastasis was reported once in an NSCLC patient with T790M EGFR tumor mutation.19 Our series consisted of patients treated before first-line setting registration of osimertinib, though four out of seven received osimertinib as post-front-line treatment. Owing to osimertinib efficacy on brain metastases demonstrated in large Phase 3 trials in patients with or without T790M mutation,20,21 osimertinib should also be considered the first-line treatment of choice in EGFR-mutant lung cancer patients with choroidal metastases.
The specific prognosis of EGFR-mutated lung cancer with choroidal metastases is unknown. In a study by Shah22 involving 194 lung cancers (NSCLC and SCLC) with uveal metastatic tumors, the 1-year mortality from eye diagnosis time was estimated at 54%, with a median survival < 12 months. In our series, the median OS in patients with choroidal metastases was 23.4 months (95%CI [0.1–51.4]) versus 27.9 months (95%CI [16.9–38.9]) in those without choroidal metastases, in line with the results of both first-line osimertinib Phase 3 trial21, in Stage IV patients with classical EGFR-activating mutations, and the second-line osimertinib AURA3 Phase 3 in patients with T790M resistance mutations20. Therefore, the specific prognostic role of choroidal metastases remains questionable.
In addition to its single-center nature, a major limitation of our study is its small sample size, accounting for its low statistical power and unmeasured confounders. However, as exon 19 mutations were an independent factor of OS in our cohort, it is probably representative of a Caucasian EGFR-mutant population.
It must be mentioned that choroidal metastasis is a rare clinical condition, encountered in 1–8% of the whole EGFR-mutant NSCLC population. We feel that our observations could support larger-scale prospective surveys on a national or international scale.