2.1 Clinical characteristics and imaging features of patients with NMOSD-NOSIS
We identified 43 (9.13%) patients with NMOSD-NOSIS among 471 patients with NMOSD (Figure 1). Table 1 summarizes the demographic and clinical features of the patients with NMOSD-NOSIS. Thirty-eight (88.37%) of these patients developed ON/MY during the follow-up, while the other five patients showed only non-opticospinal manifestations during the follow-up period. The median time within which NMOSD-NOSIS patients developed ON/MY was 12 (IQR: 8 - 26) months. The median age at onset in the NMOSD-NOSIS cohort was 28 (IQR: 21 - 37) years. The median follow-up period at the last follow-up was 5 (IQR: 3 - 8) years. The serum AQP4-IgG positivity rate was 83.72%. Serum Myelin Oligodendrocyte Glycoprotein-IgG was assayed in four out of seven patients with negative serum AQP4-IgG, and all were negative. The follow-up period of NMOSD-NOSIS patients who demonstrated only non-opticospinal manifestations was shorter than that of NMOSD-NOSIS patients who developed ON/MY (3.00 (IQR: 1.00 - 5.00) vs. 5.50 (IQR: 3.00 - 8.25), P = 0.042). There were no significant differences in other clinical manifestations and disabilities between the NMOSD-NOSIS patients who developed ON/MY and the NMOSD-NOSIS patients who showed only non-opticospinal manifestations during the follow-up period (Table 1). Among these NMOSD-NOSIS patients, 15 with vomiting or hiccups as initial symptoms first visited the Department of Gastroenterology, one with peripheral facial paralysis as the initial symptom was first diagnosed with facial neuritis, while two cases with dysphagia and quadriplegia were first diagnosed with Guillain-Barre Syndrome (GBS). Eighty-one non-opticospinal initial symptoms in these 43 NMOSD-NOSIS patients were analyzed. The most common non-opticospinal initial symptom was vomiting (67.44%), followed by hiccups (44.19%), vertigo (16.28%), diplopia (16.28%), limb weakness/numbness (9.30%), dysphagia (6.98%), somnolence (4.65%), ataxia (4.65%), headache (4.65%), facial paralysis (4.65%), facial hypoesthesia (2.33%), limb spasm (2.33%), tinnitus (2.33%), and psychiatric symptoms (2.33%) (Figure 2A). Patients with somnolence did not fulfill the criteria for narcolepsy, which has been described as a symptom of NMOSD. Fifty-seven non-opticospinal initial symptoms went untreated or were treated symptomatically, and 24 non-opticospinal initial symptoms were treated with high-dose intravenous corticosteroid therapy. Most of these initial symptoms resolved completely, even without steroid treatment in some cases. The total complete remission (CR) rate of these initial non-opticospinal symptoms was 66.67%, including a vomiting CR rate of 89.66% and a hiccup CR rate of 84.21% (Figure 2B).
To compare the difference in the recovery of non-opticospinal and opticospinal symptoms, we selected all the isolated ON relapses (n = 21) and MY relapses (n = 25) in these 43 NMOSD-NOSIS patients as controls. We found that the CR rate of non-opticospinal symptoms was higher than that of opticospinal symptoms (Figure 2B).
Among these 33 NMOSD-NOSIS patients, 29 underwent brain MRI examinations at the time of onset, 12 had brain MRI examination later during follow-up, and other two had no brain MRI examination during the disease course. All of the 12 patients undergoing brain MRI examination later during follow-up had vomiting or hiccups as initial symptoms, and no obvious abnormality was found in their brain MRI. Vomiting or hiccups in four patients, headache in two and limb spasm in one were not been verified by corresponding MRI-documented lesions at the time of onset. One patient with vomiting or hiccups as initial symptoms had no obvious cerebral lesions on the brain MRI at the time of onset, but developed a lesion in the area postrema after 6 months. Other specified clinical symptoms in those 29 NMOSD-NOSIS patients undergoing brain MRI examination at the time of onset were verified by corresponding MRI-documented lesions. The main responsible lesions for these symptoms were in the area postrema (corresponding to vomiting, hiccup, vertigo, or diplopia; 44.83%), brainstem (other areas) (corresponding to vomiting, hiccup, vertigo, dsyphagia, or limb weakness/numbness; 20.69%), parabrachialis (corresponding to vertigo, diplopia, facial paralysis, facial hypoesthesia, or tinnitus; 13.79%), brainstem (periependymal) (corresponding to hiccup, vertigo, or diplopia; 10.34%), diencephalon (corresponding to somnolence or psychiatric symptom; 10.34%), and cerebellum (corresponding to ataxia; 3.45%) (Figure 3A). Most of the lesions around the third/fourth ventricle disappeared during follow-up (disappearance rate of lesions: area postrema: 69.23%; diencephalon: 100%). The further away from the third/fourth ventricle the lesion occurred, the less likely the lesions were to disappear [The percentage of patients whose lesions disappeared during an average follow-up of 17(IQR:9-46) months in different areas is as follows: periependymal areas in brainstem: 50%; other areas of the brainstem: 66.67%; parabrachialis: 0%; and cerebellum: 0%] (Figure 3B).
2.2 Comparison of clinical characteristics, imaging features and long-term clinical outcomes between NMOSD-NOSIS and NMOSD-OSIS patients
To prevent the interference of antibody serotypes, we only compared the differences between the NMOSD-NOSIS and NMOSD-OSIS patients who were serum-positive for AQP4 antibodies. As shown in Table 2 and Figure 4, the NMOSD-NOSIS patients had a younger onset age [25(IQR: 20 - 33) years vs. 36 (IQR: 25 - 48) years, P < 0.001] and lower serum AQP4 titers than the NMOSD-OSIS patients [Log(serum AQP4 titers): 1.59 ± 0.50 vs. 1.81 ± 0.48, P = 0.030] (Figure 4A). The NMOSD-NOSIS patients had lower EDSS scores at onset and follow-up than the NMOSD-OSIS patients [EDSS at onset: 0 (IQR: 0 - 2.00) vs. 3.00 (IQR: 3.00 - 4.50), P < 0.001; EDSS at follow-up: 2.50 (IQR:1.00 - 3.00) vs. 3.00 (IQR: 2.00 - 4.00), P = 0.036]. Moreover, the NMOSD-NOSIS patients reached an EDSS of 3.0 later than did the NMOSD-OSIS patients [2.00 (IQR: 1.00 - 4.00) years vs. 1.00(IQR: 0 - 4.00) years, P = 0.018] (Figure 4B). After follow-up, there was no difference in the frequency of the ON+MY phenotype between the NMOSD-NOSIS patients and the NMOSD-OSIS patients (57.58% vs. 61.42%, P = 0.664). When analyzing all 1243 relapse episodes in the two groups, we observed that the NMOSD-NOSIS patients had more frequent isolated brain/brainstem attacks during the disease course than did the NMOSD-OSIS patients (12.40% vs. 3.11%, P < 0.001) (Table 2).
Of the total 414 patients, 18 patients did not have brain or spinal cord MRI data. Three hundred and six patients had baseline brain or spinal cord MRI data, and 163 patients underwent a brain or spinal cord MRI reexamination during the follow-up. All lesions detected in baseline and follow-up MRI examinations were recorded and included in the analysis. Brainstem lesions were more commonly present in NMOSD-NOSIS patients than in NMOSD-OSIS patients [Brainstem (area postrema): 41.94% vs. 19.11%, P = 0.004; Brainstem (periependymal): 16.13% vs. 4.47%, P = 0.027; Brainstem (others): 51.61% vs. 33.33%, P = 0.045]. NMOSD-NOSIS patients had less frequent longitudinally extensive transverse myelitis (LETM), were more likely to have cervical spinal cord lesions, and were less likely to have thoracic spinal cord lesions than were NMOSD-OSIS patients (LETM: 54.84% vs. 73.33%, P = 0.028; cervical segment: 45.16% vs. 24.35%, P = 0.011; thoracic segment: 3.23% vs. 17.97%, P = 0.035) (Table 2).
Evaluation of the pre- and post-treatment ARRs for the main treatment modalities between NMOSD-NOSIS and NMOSD-OSIS patients is shown in Table 3. Three hundred and three out of 414 patients treated with immunosuppressive agents for more than 1 year were included in the analysis; 178 were administered azathioprine (2 mg/kg daily), 105 were given mycophenolate mofetil (1 g/day), and 20 were given rituximab [500 mg through intravenous infusion, repeated 2 weeks later. This cycle was repeated every 6 months (4 cycles in total for each patient)]. ARR decreased significantly after the three major immunosuppressant therapies in both the NMOSD-OSIS and NMOSD-NOSIS groups.