Patient Demographics and Characteristics in the whole study population (Table 1)
Table 1
Baseline Characteristics of Patients with Non-Small-Cell Lung Cancer EGFR Mutation by Choroidal Metastasis Status
Variables | All n = 83 | Choroidal metastases n = 7 | Non-choroidal metastases p-value n = 76 |
Age, years Mean Median (IQR) | 69 69 (60–80) | 56 57 (50–60) | < 10− 6 70 71 (61–80) |
Gender (%) Male Female | 37 (45) 46 (55) | 2 (29) 5 (71) | 0.45 35 (46) 41 (54) |
Smoking status (%) Current smoker Former smoker Never smoker | 4 (5) 35 (42) 44 (53) | 0 (0) 4 (57) 3 (43) | 0.79 4 (5) 31 (41) 41 (54) |
Subtype of EGFR mutations (%) Exon 19 Exon 21 Others | 37 (44) 33 (40) 13 (16) | 3 (43) 3 (43) 1 (14) | 1 34 (45) 30 (39) 12 (16) |
Co-mutation status (%) Yes No Unknown | 52 (62) 29 (35) 2 (2) | 5 (71) 2 (29) 0 | 1 47 (62) 27 (35) 2 (3) |
Complex EGFR mutations (%) Yes No Unknown | 11 (13) 70 (84) 2 (2) | 2 (29) 5 (71) 0 | 0.36 9 (12) 65 (85) 2 (3) |
TP53 co-mutation (%) Yes No Unknown | 40 (48) 41 (50) 2 (2) | 4 (57) 3 (43) 0 | 0.76 36 (47) 38 (50) 2 (3) |
IQR, interquartile range |
EGFR, epidermal growth factor receptor. |
We identified 83 cases of metastatic NSCLC harboring an EGFR mutation (Table 1) in our electronic files during a three-year period. Of these, 20 cases were diagnosed earlier but referred to our Department for treatment during this time period. The other 63 were newly diagnosed NSCLC cases. Main clinical baseline characteristics are summarize in Table 1.
Among the 83 EGFR-mutant NSCLC patients, six exhibited choroidal metastases, revealed by visual disturbances. An additional case (Fig. 1) was identified upon a retrospective systematical analysis of all brain MRIs performed by an expert radiologist, resulting in a final number of 7 choroidal metastasis patients (8.4%). This latter patient had no clinical visual disturbances.
Tumor Genetic Profile in the whole study population (Table 1)
Thirty-seven (44%) tumors harbored exon 19 mutations (either 32 (38%) deletions, and 5 (6%) deletion–insertions). Overall, 33 (40%) tumors had a mutation in EGFR exon 21, (31 [38%] L858R, and two [2%] L861Q). Four (5%) tumors exhibited a missense mutation in exon 18, and nine (11%) a mutation in exon 20.
We found 11 tumors (13%) with EGFR complex mutations associated with common or uncommon mutations, and two had two rare mutations associated together. Two tumors (2%) presented with a T790M mutation in exon 20 of the EGFR gene at diagnosis, associated with a common activating mutation. Co-mutations involving the TP53 and EGFR genes were found in 40 tumors (48%).
Furthermore, 4/83 (5%) patients experienced SCLC transformation during disease evolution, involving 3/76 (4%) patients without choroidal metastasis, and 1/7 (14%) with choroidal metastasis (Patient #5). These four tumors harbored TP53 mutation at diagnosis.
Case Reports of One Typical Patient with Choroidal Metastasis
Case of Patient #1. A 60-year-old man, 12 pack-years having quitted smoking for 18 years, reported right visual acuity loss as well as asthenia and coughing. A fundus examination revealed a well-defined yellowish-colored circular sub-macular lesion on the right eye (Fig. 2A). Initial B-scan echography demonstrated a right dome-shaped lesion on the macula with hyper-reflectivity and maximal elevation of 1.12 mm (Fig. 2B). A hyper-fluorescent choroidal macular right lesion was evidenced via fluorescein angiography (Fig. 2C, left); optimal coherence tomography showed a retinal detachment with sub-retinal fluid on the right eye (Fig. 2B, right). As a choroidal metastasis was suspected, a whole-body computed tomography (CT) scan revealed a left upper lobe (LUL) pulmonary mass. A positron emission tomography-CT (PET-CT) showed high fluorodeoxyglucose (FDG) uptake on the pulmonary mass, along with a vertebral bone metastasis. A TTF1-positive, ALK and ROS-1 negative, Rb positive adenocarcinoma with a PD-L1 tumor proportion score (TPS) below 1% was diagnosed in July 2018. A molecular analysis on the tumor sample revealed a deletion in exon 19 of the EGFR gene (c.2236_2250 del).
Afatinib was initiated as first-line treatment in August 2018. Within a month of therapy, the patient reported a complete resolution of his visual symptoms and chest CT-scan showed partial response. After 9 months of therapy, a thoracic disease progression occurred without recurrence of ocular symptoms or new choroidal metastasis, upon ophthalmological examination. Starting in May 2019, he received four cycles of platinum salt and pemetrexed and achieved stable disease. Maintenance therapy with pemetrexed alone began in August 2019. Two months later, the patient reported a relapse of the right eye visual impairment. Ophthalmological examinations showed an irregular thickening of the macula with serous retinal detachment, highly suggestive of a chorioretinal scar rather than a new progressive cancer lesion.
Clinical and Molecular Characteristics of Choroidal Metastasis Patients (Table 2 and Table 3)
Table 2
Clinical and Molecular Characteristics of Seven Patients With Choroidal Metastases
Patients | Patient #1 | Patient #2 | Patient #3 | Patient #4 | Patient #5 | Patient #6 | Patient #7 |
Sex | Male | Female | Female | Female | Female | Female | Male |
Ethnic group | Moroccan | Caucasian | Caucasian | Caucasian | Chinese | Guinean | Moroccan |
Age-ranges at the time of diagnosis | 60–65 years | 60–65 years | 45–50 years | 50–55 years | 45–50 years | 50–55 years | 50–55 years |
History of smoking | 12 pack-years, quitted 18 years ago | 27 pack-years, quitted 27 years ago | No | 20 pack-years (active) | No | No | 30 pack-years (active) |
TNM stage at time of diagnosis (8th classification) | cT3N0M1c | cT1cN3M1c | cT4N3M1c | cT4N3M1c | cT2N3M1c | cT3N2M1c | cT4N2M1c |
EGFR mutation(s) at time of diagnosis | Deletion of exon 19 c2236_2250del | Deletion of exon 19 c2240_2257del | Complex mutation of exon 21 Leu833Val His835Leu | Mutation of exon 21 L858R Mutation of exon 20 T790M (AF = 11.6%) | Mutation of exon 21 L858R | Mutation of exon 21 Leu858R | Deletion of exon 19 c2235_2249del |
Mutation TP53 | No | Exon 5 | No | Exon 5 | Exon 7 | No | Exon 7 |
IHC Rb | Positive | Negative | Positive | Positive | Negative | Negative | Negative |
Treatment (number of lines) | 2 | 5 | 8 | 5 | 4 | 2 | 4 |
Anti-EGFR treatments (number of lines) | 1 | 3 | 2 | 3 | 1 | 1 | 3 |
Anti-EGFR TKIs | Afatinib L1 | Afatinib L1 Osimertinib L3 Osimertinib L5 | Afatinib L6 Osimertinib L7 | Erlotinib L1 Osimertinib L2 Osimertinib L5 | Erlotinib L1 | Erlotinib L2 | Erlotinib L1 Osimertinib L2 Afatinib L4 |
Carcinomatous meningitis during disease course | No | Confirmed by cytology | High MRI probability | High MRI probability | No | Confirmed by cytology | High MRI probability |
Transformation in SCLC during disease course | No | No | No | No | Yes | No | No |
Alive on Aug. 2020 | Yes | No | No | No | No | No | No |
Follow-up from diagnosis | 25 months | Death at 17 months | Death at 23 months | Death at 27 months | Death at 6 months | Death at 2 months | Death at 36 months |
EGFR, epidermal growth factor receptor; IHC, immunohistochemistry; TKI, tyrosine kinase inhibitor; SCLC, small cell lung cancer; MRI, magnetic resonance imaging; AF, allelic frequency. |
Table 3
Diagnosis Characteristics of Choroidal Metastases in Seven Patients
Patients | Patient #1 | Patient #2 | Patient #3 | Patient #4 | Patient #5 | Patient #6 | Patient #7 |
Symptoms | Yes | Yes | Yes | Yes | Yes | Yes | No |
Time to onset of symptoms (months) | 0 | 0 | 0 | 0 | 0 | 0 | |
Nature of symptoms | Myodesopsia and visual acuity loss of right eye | Intermittent visual blur on right eye | Visual acuity loss of left eye | Visual acuity loss of right eye | Blurred vision in right eye | Retro-ocular pain and visual acuity loss on right eye | |
Associated metastatic site (yes / no) | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
Associated brain metastases | No | Yes | No | Yes | Yes | Yes | Yes |
Diagnostic tool | Fundoscopic examination Fluorescein angiography B-scan ultrasound | Fundoscopic examination Fluorescein angiography B-scan ultrasound | B-scan ultrasound | Fundoscopic examination Fluorescein angiography B-scan ultrasound | Fundoscopic examination Fluorescein angiography B-scan ultrasound | B-scan ultrasound | Retrospective MRI analysis |
Uni- or bilateral lesion | Unilateral | Unilateral | Unilateral | Unilateral | Bilateral | Unilateral | Unilateral |
Visibility on MRI (yes / no) | Not performed | No | Yes | Yes | Yes | Yes | Yes |
Response to systemic therapy | Objective response to afatinib | Objective response to afatinib | Objective response to chemotherapy | Objective response to osimertinib | No | Not evaluable | Objective response to erlotinib |
Radiotherapy of choroidal metastases | No | No | No | No | Yes | No | No |
Concomitant progression to visceral progression | No | No | No | No | No | Not evaluable | No |
MRI, magnetic resonance imaging. |
Seven patients (8.4%) with choroidal metastases associated with EGFR-mutant NSCLC were retrieved in our retrospective cohort. Their clinical characteristics at diagnosis are shown in Table 2, while the choroidal metastasis features and their evolution are summarized in Table 3.
Patient Demographics and Characteristics of Choroidal Metastasis Patients
When comparing the clinical characteristics of the choroidal metastasis group with the remaining 76 patients, no statistically significant differences were found, except for a lower mean age in the choroidal metastasis group, (p < 10− 6). Most of the choroidal metastasis patients were women (five/seven cases, 71%), and the majority (four of seven cases, 57%) were current or former smokers, with two being active smokers. There was only one Asian patient, with three being of African origin.
In all patients (except Patient #7, with retrospective MRI-based diagnosis), the choroidal metastases resulted in symptoms, such as pain, blurred vision, or visual acuity loss. The choroidal metastases were frequently unilateral (six of seven cases, 85%). Brain MRI was performed at diagnosis in six out of seven patients, with choroidal metastases actually visible on MRI after a retrospective secondary analysis by an expert radiologist in five out of six patients, though all these lesions were missed on the first MRI report, except for those of Patient #6.
The choroid metastases were associated with other metastatic sites like brain metastases in five out of seven patients without any obvious leptomeningeal carcinomatosis on brain MRI. Four out off five evaluable patients who were treated with TKI showed an objective response with visual symptom improvement. Local response was maintained in all cases, and no choroidal local progression occurred, even in case of tumor progression at other metastatic sites. Four out of seven patients received osimertinib as second-line therapy based on the emergence of T790M resistance mutation. All patients received a pemetrexed/platinum-based chemotherapy doublet during the course of the disease. One patient out of seven also underwent radiotherapy of the choroidal metastasis.
Genetic Profile of EGFR-Mutated NSCLC of Choroidal Metastasis Patients
The molecular tumor status of the choroidal metastasis patients was consistent with that of the remaining population, as 85% had a common activating EGFR mutation (Table 1). One patient had a rare, complex exon 21 mutation diagnosed late in the NSCLC course, using whole exome sequencing.
One patient displayed a T790M mutation in the EGFR gene exon 20 at diagnosis at the same allele frequency as the common exon 21 mutation (Patient #4, L858R mutation). The tumor molecular patterns of the seven patients with choroidal metastases are shown in Table 2: four patients had TP53 mutations, four had lost Rb immunohistochemical expression, indicative of a Rb probable gene deletion, and three had simultaneously both events at diagnosis supporting the preexistence of a sub-clone with small-cell carcinoma features. However, during the study period of time, only one case of SCLC histological transformation was observed in one out of seven patients (Patient #5), who similarly exhibited TP53 mutation and Rb loss expression in the diagnostic sample.
Overall Survival Analysis
Median follow-up time in the whole study population was 42.2 mo., 95%CI [37.2–47.1]. On the study termination date, 28 patients (33.7%) were still alive and 55 patients (66.3%) were deceased. No patients were lost to follow-up.
Median OS in the whole study population was 27.9 mo. (95% CI [17.1–38.6]) with 55.3% and 24.1%, 2-year and 5-year OS, respectively.
In univariable analysis, median OS in the choroidal metastasis group was 23.4 mo. 95%CI [0.1–51.4]) versus 27.9 mo. 95%CI [16.9–38.9] in the non-choroidal metastasis group (p = 0.32) (Fig. 3). In the choroidal metastasis group, 2-year and 5-year OS were 47.6% and 0%, respectively. In the non-choroidal metastasis group, 2-year and 5-year OS were found to be 55.8% and 26.3%, respectively. The presence of EGFR exon 19 deletion was the only variable significantly associated with a longer median OS (52.6 mo., 95%CI [24.1–81.1]) as compared with patients showing EGFR exon 21 point mutation (26.6 mo., 95%CI [15.3–37.9]) or non-conventional EGFR mutations (19.8 mo., 95%CI [15.3–24.4]), p = 0.03.
After adjusting for variables showing a p-value ≤ 0.02 in the univariate analysis, i.e. EGFR mutation subtype, and TP53 mutations, choroidal metastasis did not show any impact on survival (Table 4) while EGFR mutation subtype remained as an independent prognostic factor, since patients with exon 21 and uncommon EGFR mutations had significantly shorter survivals, with adj. HR = 1.89, 95%CI [1.01–3.55] (p < 0.05), and 2.42 95%CI [1.11–5.31] (p = 0.03), respectively, as compared with patients whose tumor harbored EGFR exon 19 deletion.