We identified consecutive patients with aphasic MCI-LB who were prospectively recruited at Nippon Life Hospital, Osaka, Japan, between September 2018 and August 2021. The inclusion criteria were as follows: (1) diagnosis of aphasic MCI and (2) diagnosis of probable MCI-LB (2). The exclusion criteria were as follows: (1) significant impairment in cognitive domains other than language; (2) history of neurological diseases other than prodromal DLB, psychiatric diseases, or hearing impairment; and (3) evidence of focal brain lesions other than atrophy on magnetic resonance imaging. All patients underwent standard neuropsychological, speech, and language assessments by an experienced speech and language therapist and clinical neuropsychologist and routine laboratory investigations. N-isopropyl-p-[¹²³I] iodoamphetamine single-photon emission computed tomography (IMP-SPECT) were performed. The diagnostic classification was blinded to the results of the imaging analysis and was based on the results and recordings of the speech and language assessments and samples of general conversation, which were reviewed in a consensus meeting 2–4 weeks after the clinical assessments.

All patients and their caregivers provided written informed consent. This study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Ethics Committee of Nippon Life Hospital.

To determine the features of cognitive and behavioral alterations, all patients underwent the following tests: Clinical Dementia Rating (CDR) (22); Mini Mental State Examination (MMSE); Addenbrooke's Cognitive Examination-Revised (ACE-R) (23, 24); Alzheimer's Disease Assessment Scale (ADAS) (25); digit span and spatial span tests; and noise pareidolia tests (26, 27), in which visual hallucination-like illusions were evoked and measured.

The following questionnaires were administered to the patients' caregivers: the Neuropsychiatric Inventory (28) to assess hallucinations and sleep disturbances and the Cognitive Fluctuation Inventory (29) to assess fluctuations in cognition.

After bedside language assessments (30), the following in-depth evaluations were conducted: the Western Aphasia Battery (WAB) (31, 32), which served as the primary measure of general language ability; and the Test of Lexical Processing in Aphasia (TLPA) (33). The TLPA is a standardized, widely used language test for Japanese speakers, and a total of 200 items in line-drawing cards are included in the picture-naming task or the auditory word comprehension task. In the TLPA-word comprehension task, after listening to a spoken word, patients are asked to match one of 10 line-drawing cards. The TLPA-word comprehension task was used to classify the severity of single-word comprehension impairment. TLPA-word comprehension scores were categorized as follows: normal, 198–200; minimal impairment, 181–197; mild impairment: 161–180; severe impairment, <161 (6).

The extent of hypoperfusion on IMP-SPECT was determined by three-dimensional (3D) stereotactic surface projections (SSPs) (34), in which IMP-SPECT images were sampled at 16,000 predefined cortical locations and projected on the 3D image after realignment, spatial normalization, and non-linear warping. The voxel values of an individual's IMP-SPECT data were normalized to the whole-brain tracer uptake and compared with an age-matched normal database, yielding a 3D SSP Z-score image; the abnormalities of cerebral hypoperfusion were displayed with a Z-score map. Z-scores were calculated using the following equation: Z-score = [(control mean)–(individual value)]/(control standard deviation). We used a Z-score of 2 as the cut-off value in each voxel, and voxels with a Z-score of ≤2 were considered voxels without significantly decreased regional cerebral blood flow.

Eight patients with both aphasic MCI and probable MCI-LB were enrolled in this study (Figure 1). During the study enrollment period, 26 patients were diagnosed with aphasic MCI at the Brain Function Center of Nippon Life Hospital. Of the 26 patients, 20 met the core criteria for PPA. Of the 20 patients with PPA, three met the criteria for nfvPPA, two for svPPA, five for lvPPA, and 10 for the unclassified fluent variant of PPA (primary progressive anomic aphasia, n = 8; primary progressive TCSA, n = 2). Moreover, of the 20 patients with PPA, two met the criteria for probable MCI-LB (lvPPA, n = 1; primary progressive anomic aphasia, n = 1). Of the remaining six patients who did not meet the core criteria for PPA, two had logopenic progressive aphasia, and four had progressive anomic aphasia. Thus, of the 26 patients with aphasic MCI, eight (30.8%) were finally included in the study because of a diagnosis of probable MCI-LB (logopenic: n = 3; anomic: n = 5). One patient, whose case has been published as a single case report (16), was not enrolled in the current study. For clarity, case numbers were arranged in order of the WAB-Aphasia quotient score.

Three patients were women, and all but one patient (case 5) were right-handed (Table 1). The median age of onset was 69.5 years (62–81 years) [median (range)]. Two patients experienced symptom onset before 65 years of age. All eight patients had experienced language impairment for ≤5 years; of these, the majority (5/8) had experienced language impairment for ≤3 years. The results of routine laboratory tests were unremarkable. None of the patients had a family history of neurological disease.

All patients met the criteria for probable MCI-LB (Table 1) (2). None of the patients showed significant generalized dementia (CDR score: 1–3), as measured by the CDR at visit 1. Episodic memory for daily events was relatively well preserved in all patients compared with language function. However, the majority showed scores on the ADAS word recall subtest below the lower limit of healthy individuals. Similarly, three patients (cases 6–8) showed total scores on the MMSE and/or ACE-R below the lower limit of healthy individuals. All these general cognitive tasks, including memory tasks, more or less rely on language function, and the scores are affected by aphasia, in particular word-finding difficulty. However, all but one patient (case 8) did not show a score >2 SD below the mean on the ADAS word recognition subtest, which is considered to be less affected by word-finding difficulties than the ADAS word recall subtest. These findings indicated that general cognitive function except for language function was well preserved in all but one patient (case 8). Case 8 had mild episodic memory and visuospatial impairments. However, his language impairment was more prominently severe than other cognitive impairments, and was the primary cause of impaired activities of daily living (5).

Regarding the core clinical features of MCI-LB (2), three patients (37.5%) exhibited visual hallucinations, seven patients (87.5%) had Parkinsonism, seven patients (87.5%) had fluctuating cognition, and five patients (62.5%) exhibited rapid eye movement sleep behavior disorder.

Regarding the proposed biomarkers of MCI-LB (2), three patients were assessed by dopamine transporter (DAT) SPECT imaging, and all three showed reduced DAT uptake. Two patients were assessed by ¹²³I-metaiodobenzylguanidine (¹²³I-MIBG) myocardial scintigraphy, and both showed decreased cardiac MIBG uptake. Additionally, two patients were assessed by both DAT-SPECT imaging and ¹²³I-MIBG myocardial scintigraphy, and both showed reduced DAT uptake and decreased cardiac MIBG uptake. The remaining patient (case 5) was not assessed by DAT-SPECT imaging or ¹²³I-MIBG myocardial scintigraphy but showed all core clinical features of MCI-LB (2).

At visit 1, two patients (cases 4 and 8) showed no prominent cognitive, behavioral, or psychiatric symptoms, except for progressive aphasia but reported experiencing the core clinical features of MCI-LB at visit 2.

The language features are summarized in Table 2. At visit 1, the median WAB-Aphasia quotient score was 85.4 (74.2–93.6). All patients had fluent spontaneous speech, and there was no evidence of typical nfvPPA features (5), including dysprosody, apraxia of speech, and agrammatism. In addition, none of the patients showed typical svPPA features (5) that reflect a loss of semantic knowledge, such as poor word comprehension with an obvious loss of single-word meaning, surface dyslexia/dysgraphia, visual impairments in face and object recognition, or object misuse (4, 35–40). Three patients (cases 6–8) showed typical lvPPA features (5), including anomia, phonemic paraphasia, and impaired repetition of polysyllabic words or phrases. Thus, they were diagnosed with logopenic progressive aphasia. However, the remaining five patients (cases 1–5) did not show the typical features of nfvPPA, svPPA, or lvPPA (5). Their single-word comprehension ability was within the normal/minimal range of the TLPA-word comprehension task. Moreover, repetition of polysyllabic words or phrases formed by up to four words of 14 morae on the WAB-Repetition task (“Nihon Koko yakyu remmei” “Japan High School Baseball Federation”) was preserved. Thus, the severity of the repetition deficit was too subtle as the typical lvPPA feature (6, 41). In addition, none of these patients showed phonemic paraphasia. Their language features were characterized by apparent anomia, and their repetition and comprehension abilities were relatively preserved. Therefore, based on the current research findings for unclassified fluent variants of PPA (6, 42, 43), five patients (cases 1–5) were diagnosed with progressive anomic aphasia.

At visit 2, three patients (cases 2, 4, and 8) underwent follow-up examinations after 1 year (Table 2). None of the three patients showed non-fluent features such as dysprosody, apraxia of speech, or agrammatism (5). In case 4, word comprehension impairment in the TLPA-word comprehension task drastically progressed to severe at visit 2. In contrast, the severity of the repetition deficit was mild compared to that of typical lvPPA (6, 41) because the repetition ability of polysyllabic words or phrases formed by up to four words of 14 morae on the WAB-Repetition task was preserved. Thus, the pattern of aphasia at visit 2 changed from progressive anomic aphasia to progressive TCSA, characterized by relatively preserved repetition but poor word comprehension (6). In contrast with case 4, the word comprehension ability of case 2 in the TLPA-word comprehension task was within the normal range, but the repetition ability of polysyllabic words or phrases in the WAB-Repetition task deteriorated at visit 2. Thus, the pattern of aphasia changed from progressive anomic aphasia to logopenic progressive aphasia at visit 2. In case 8, general language function, including word comprehension ability for familiar/frequent words, drastically deteriorated at visit 2. Thus, the pattern of aphasia at visit 2 changed from logopenic progressive aphasia to progressive Wernicke's aphasia, characterized by poor repetition and word comprehension (6). Case 8 also showed extensive and severe deficits in cognitive domains other than language.

The IMP-SPECT-3D SSP analysis of each patient is shown in Figure 2. The degree of hypoperfusion varied between patients; when present, it most often involved the lateral or medial occipital lobe (seven patients: cases 1–4, and 6–8), revealing potential biomarkers of MCI-LB (2) and supportive biomarkers of DLB (3). In addition, there was no hypoperfusion in the precuneus and posterior cingulate cortex in two out of the seven patients (cases 2 and 7), indicating the cingulate island sign, which is one of the supportive biomarkers of DLB (3). In contrast, hypoperfusion in the remaining six patients was noted in the precuneus or posterior cingulate cortex, indicating a surrogate marker of early Alzheimer's disease (44). The left temporal and parietal lobes were involved in all but one patient; hypoperfusion in the right temporoparietal lobe was present in case 5. Compared with the five patients (cases 1–5) with progressive anomic aphasia, the three patients (cases 6–8) with logopenic progressive aphasia had extensive hypoperfusion in the left temporoparietal cortices, including Wernicke's area.

Three patients (cases 1, 4, and 8) underwent cholinesterase inhibitor therapy with donepezil, and all patients showed improvement in general cognitive functions, including language function. Case 8 had rapid temporary improvement in language and other cognitive functions, based on bedside examinations by a behavioral neurologist and speech and language therapist. The pattern of aphasia in case 8 improved from progressive Wernicke's aphasia at visit 2 to a relatively similar pattern of logopenic progressive aphasia observed at visit 1. However, the effect lasted for only 1 week.

The remaining two patients (cases 1 and 4) also underwent language and neuropsychological assessments, and we compared their language and cognitive impairments at baseline with their performance at follow-up sessions after donepezil treatment. Follow-up started 3 months after the donepezil dosage was increased to 5 mg/day (16); we observed improvements in the patients' total WAB-Aphasia quotient scores and TLPA-naming tasks, and in the total MMSE and ACE-R scores (Table 3).