Maximizing the benefits of disease-modifying treatments (DMTs) for Alzheimer’s disease (AD) requires early identification of cognitive impairment and abnormal brain amyloid-beta (Aβ) status. Either one alone is insufficient. Additionally, clinical trials of DMTs are impeded by high screen failure rates and costly prescreening. Thus, an efficient and cost-effective solution to streamline the process of early AD identification is urgently needed. This study aimed to assess the accuracy of a brief digital cognitive assessment, the Linus Health Digital Clock and Recall (DCR), to identify cognitive impairment and predict brain Aβ status. 930 participants (mean age 72.0±6.7; 56.8% female; 23% minorities) in the Bio-Hermes-001 study were classified as cognitively unimpaired, mild cognitive impairment, or probable Alzheimer’s dementia, and 35.1% were Aβ+ on 18F-florbetapir PET scan. A DCR-based algorithm (LinusAD) used age, APOE status, drawing metrics, speech and acoustic features, and temporal-spatial features of stylus manipulation. LinusAD was compared with MMSE and blood-based biomarkers (BBMs) Lilly pTau-217, Quanterix pTau-181, C2N Aβ42/40, and C2N Amyloid Probability Score (APS). Superiority or non-inferiority was established if 95% confidence interval of the bootstrapped AUC difference between two tests was higher than or within ΔAUC±0.1. Cognitive-impairment identification by the DCR (AUC=0.85) was superior to Aβ42/40, pTau-181, and pTau-217 (AUCs=0.63, 0.66, 0.72) and non-inferior to RAVLT and MMSE (AUCs=0.89, 0.82). Aβ-status prediction by LinusAD (AUC=0.882) was equivalent to BBMs (AUC range: 0.775–0.887; average AUC=0.82) and superior to MMSE (AUC=0.71) (see Table and Figure). Combining the DCR with Aβ42/40, pTau-181, and pTau-217 improved their Aβ-status prediction performance to AUCs of 0.882, 0.835, and 0.905, respectively. The 3-minute DCR was superior to BBMs in detecting cognitive impairment and boosted their Aβ-PET prediction ability to levels comparable to CSF biomarkers of AD. Digital cognitive assessments that leverage AI process metrics, such as the DCR, enable cost-effective integration into multi-step clinical workflows to prioritize the most suitable patients for BBM testing, DMTs, and decrease high screen-failure rates in AD clinical trials.

Early detection of cognitive impairment is crucial for maximizing the benefits of disease-modifying treatments for Alzheimer’s disease (AD). Brief, automatically-scored digital cognitive assessments such as the Digital Clock and Recall (DCR) show promise in streamlining this early detection. However, wide adoption of such assessments in diverse populations requires evaluation of their demographic biases. Here, we compared the biases due to ethnicity, race, and education level between the DCR and the Mini-Mental State Examination (MMSE). We studied 706 primarily English-speaking participants from Bio-Hermes-001 study (age mean±SD = 71.5±6.7; 58.9% female; years of education mean±SD = 15.4±2.7; 85.1% White; 9.3% Hispanic), classified a priori as cognitively unimpaired (CU; n = 360), mild cognitive impairment (MCI; n = 234), or probable Alzheimer’s dementia (pAD; n = 111) based on expert consensus and neuropsychological evaluation. We also studied 770 participants in a prescreening study for AD-related clinical trials (age mean±SD = 68.8±13.2; 64% female; education = 15±2.6; 78% White; 20% Hispanic) including CU (n = 338), MCI (n = 178), and pAD (N = 254). For each dataset, bias was compared by bootstrapping two multiple linear regressions predicting either Z-scaled DCR or MMSE scores using race, ethnicity, sex, education, and age as predictors. The bootstrapped demographic coefficients (i.e. mean differences) were compared between tests and significance was determined via 95% confidence intervals on these differences. A larger ethnicity bias was observed for the MMSE than for the DCR, which was significant in the Bio-Hermes (bias difference = 0.44 larger for MMSE, 95% CI = 0.12–0.75) but not the pre-screener study (bias difference = 0.20 larger for MMSE, 95% CI = -0.32–0.74). We found significant differences between Hispanic and non-Hispanic individuals only for MMSE in both datasets (Mann-Whitney test, p <0.01). Bias for race and education was not significantly different between tests. Unlike the MMSE, the DCR is not influenced by ethnicity. Given the higher prevalence of cognitive impairment and greater risk of dementia in ethnic minorities (with Hispanics forming the largest minority group), the deployment of less biased assessments such as the DCR is important for more equitable cognitive screening.

Disease-modifying treatments for Alzheimer’s disease highlight the need for early detection of cognitive decline. However, most primary care providers do not currently perform routine cognitive testing, in part due to a lack of time and resources to administer and interpret the tests. Brief, self-scoring, and sensitive digital cognitive assessments, such as the Linus Health Core Cognitive Evaluation (CCE)–which includes the Digital Clock and Recall (DCR™) and the Life and Health Questionnaire (LHQ)–can automatically provide medically-informed recommendations that can address this need. Here we evaluate the clinical appropriateness of the recommendations generated by this clinical decision support (CDS) tool to guide the diagnosis of cognitive impairment by primary care providers (PCPs). The CDS tool uses data from the CCE to list potential medical concerns and recommend pathways toward formal diagnosis and/or care. We conducted a retrospective expert-review study in June 2023 to evaluate the nine CDS pathways for patients aged 55 and above. Experts were five board-certified cognitive neurologists affiliated with academic institutions. We calculated the median ratings (on a scale of 1 to 9, where 9 is high appropriateness) for the nine CDS pathways across raters. A rating of 7 or above was deemed clinically appropriate. All 7 pathways related to cognitive impairment received a clinically appropriate rating (median = 7, SD = 0.3, range=7-8). Pathways below the appropriateness threshold included the one for Green DCR scores (i.e., cognitively unimpaired; median = 6, SD = 0.87) and a preliminary Lecanemab eligibility pathway (median = 5, SD = 1.10). Pathways and the cognition-related recommendations generated by the CDS tool of the Linus Health CCE are clinically appropriate, as rated by this initial survey of board-certified, academic cognitive neurologists. The findings indicate the clinical utility of digital cognitive assessments such as the CCE for guiding the PCPs’ approach to diagnosis and management of patients with cognitive impairment.

Mild cognitive impairment (MCI), is characterized by cognitive dysfunction not severe enough to affect one’s activities of daily living (ADLs)1. Annually, approximately 15-20% adults 65 and older will present with MCI 1 . MCI is considered a significant risk factor and a robust predictor for developing dementia. The time course for progression to dementia can vary substantially between individuals and is impacted by the specific pathology underlying the MCI, and the cognitive deficits associated with cognitive impairment (CI) subtypes 2,3 . Despite the conversion risk of MCI to dementia and the effectiveness of early lifestyle interventions to mitigate the conversion risk, many investigations do not account for MCI in their CI prediction models. This research investigates binary and 3-class ML-enabled modeling to classify CI status leveraging multiple modalities of cognition extracted from the Digital Clock and Recall (DCR), a brief digital cognitive assessment. Data from 983 participants in the Bio-Hermes-001 multi-site study (age mean±SD=72±6.7; 56% female; years of education mean±SD=15±2.7; primary language English), a priori classified as cognitively unimpaired (CU; n=417), mild cognitively impaired (n=309), or probable Alzheimer’s dementia (n=257) based on expert consensus clinical diagnosis and neuropsychological evaluation were analyzed. A random forest model was trained on DCTclock and word recall data to classify cognitive impairment using a binary (CI and CU) and 3-tier (CI, Indeterminate, and CU) prediction thresholding schemes. The 3-tier model predictions performed well (AUC=0.887; accuracy=0.834; NPV=0.801; PPV=0.859) outperforming the binary predictions (AUC=0.865; accuracy=0.79; NPV=0.737; PPV=0.837) when measured on Biohermes’ cohort diagnosis. The sensitivity and specificity of the 3-tier predictions were 0.858 and 0.8, respectively. The DCR, a 3-minute digital cognitive assessment can be used to classify MCI and probable Alzheimer’s dementia with high accuracy, NPV, and PPV.

There is an urgent need for neuropsychological screening tests that are easily deployed and reliable. We have developed a digital neuropsychological screening protocol that is administered on a tablet, automatically scored using artificial intelligence, and requires approximately 10 minutes to administer. This tablet-administered protocol assesses the requisite neurocognitive constructs associated with emergent neurodegenerative illness The digital protocol was administered to 77 ambulatory care/ memory clinic patients (Table 1). The protocol is comprised of a 6-word version of the Philadelphia (repeatable) Verbal Learning Test [P(r)VLT], three trials of 5 digits backward (BDST), and the ‘animal’ fluency test. The protocol provides a panel of six traditional measures as would be obtained using paper/ pencil tests and manual scoring of (P[r]VLT free recall/ recognition hits, backward digit span, ‘animal’ fluency output); a variety of outcome measures quantifying errors and the process used to bring tests to fruition; and two separate, norm-referenced summary scores measuring executive control and memory. Cluster analysis using the panel of 6 traditional measures classified participants into normal (nl= 23), amnestic MCI (aMCI= 17), dysexecutive MCI (dMCI= 23), and dementia (dementia= 23) groups. Subsequent analyses of error and process variables operationally defined key features associated with amnesia including rapid forgetting such as (P[r]VLT immediate free recall trial 2 vs. delay free recall (aMCI & dementia < dMCI & nl; p< 0.001), the production of extra-list intrusion errors (dementia > nl; p< 0.002); profligate responding to recognition foils (aMCI & dementia > dMCI & nl; p< 0.001); key features underlying reduced executive measures (i.e., BDST perseveration/ related errors (dMCI & dementia > aMCI & nl, < 0.050); and the strength of semantic association from successive ‘animal’ fluency responses (nl & dMCI > dementia; p< 0.028). The novel executive and memory index scores dissociated all four groups from each other (p< 0.014). This digitally administered and scored protocol yields patterns of impaired performance similar to paper/ pencil tests. The availability of both traditional and error/ process measures suggests that subtle, nuanced indications of early emergent illness may be identified in a fast, efficient, yet comprehensive way.

A typical paper/pencil neuropsychological evaluation to assess for mild cognitive impairment (MCI) and dementia is lengthy. There is a need for a brief, digitally administered/scored neuropsychological protocol that can differentiate patients who are cognitively normal versus MCI and dementia. This need is particularly acute with the advent of disease-modifying medications to treat MCI and early Alzheimer’s disease (AD). The Digital Assessment of Cognition (DAC) assesses memory with a 6-word Philadelphia (repeatable) Verbal Learning Test [P(r)VLT]; executive abilities with three trials of 5 digits backward (BDST); and language/lexical access abilities with the ‘animal’ fluency test. 105 ambulatory care/memory clinic patients were assessed. DAC outcome measures include words recalled, recognition discriminability, BDST serial recall, and ‘animal fluency output; and separate episodic and executive summary indices. A portion of this sample (n = 70) underwent traditional, paper and pencil assessment. Cluster analysis (Table 1) classified participants into groups suggesting normal (NL = 25), dysexecutive MCI (n = 25), amnestic MCI (n = 11), mixed MCI (n = 24), and dementia (n = 25) cognitive abilities. Concordance between cluster classification and clinical diagnosis using a lengthy paper/pencil protocol was 90%. The ANOVA for the episodic memory summary index found no difference between NL/dMCI groups; all other groups differed. The ANOVA for the executive index found no differences between NL/aMCI, and dMCI/mixed MCI groups; all other groups were differentiated from each other. Within-group, dMCI patients scored lower on the executive versus the memory index (p<0.001); aMCI patients scored lower on the memory versus executive index (p<0.001); NL patients scored higher on the executive versus the memory indices (p< 0.021); and mixed MCI patients did not differ on either index. The DAC memory index was correlated with CVLT-short form delayed free recall (r = 0.648; p< 0.001) and recognition discriminability (r = 0.625, p<0.001) scores. The DAC executive index significantly correlated with WMS-IV Symbol Span (r = 0.620, p<0.001; Trails B (r = 0.477, p< 0.001), and letter fluency (r = 0.538, p<0.001). The DAC is able to identify clinically meaningful groups, showed diagnostic concordance with lengthy paper/pencil assessment, and is positively correlated with traditional paper and pencil tests. The DAC provides a reasonable means to screen for patients with putative MCI and dementia.

Amnestic and vascular dementia are two common types of dementia. Currently, diagnosing and differentiating between these two conditions requires comprehensive neuropsychological testing, neuroimaging studies, and cerebrospinal fluid analysis. Identifying these conditions at early stages, i.e., mild cognitive impairment (MCI), is crucial for maximizing the therapeutic benefits of pharmacological agents or lifestyle modifications. Here we evaluated the utility of metrics derived from the DCTclock, a brief ML-enabled digital cognitive assessment, in differentiating between amnestic and vascular MCI. We analyzed DCTclock tests from N = 215 participants (age 48-90, 126 females, education = 6-20 yrs) including patients at Lahey Hospital and Medical Center and participants in the Framingham Heart Study, for whom a diagnosis of vascular (n = 46) or amnestic (n = 169) MCI was available. Mann-Whitney U Tests were used to compare the distributions for 78 features derived from the DCTclock performance. Random forest models with recursive feature elimination and 5-fold cross-validation were used to differentiate between the two MCI subtypes. 16 DCTclock features returned p-values below the 0.05 threshold for statistical significance, and one feature reached the Bonferroni-corrected threshold of 6.4e-4. The features showing the largest difference between the two populations were the total time to complete the copy clock (p = 4.2e-4), drawing process efficiency on the copy clock (p = 1.7e-3), and average latency on the copy clock (p = 3.7e-3). 13 of the 16 top features were derived from the copy clock test. A random forest model using only 6 features (figure below) returned an accuracy of 77.9%. Metrics derived from the process of DCTclock performance, especially those obtained during the copy clock condition, showed utility for differentiating between vascular and amnestic MCI. Comparing cognitive phenotypes at the early stage of MCI is arguably a more difficult task than comparing those phenotypes among individuals who have progressed to dementia and thus exhibit more pronounced signs of cognitive impairment. More data from individuals with such cognitive profiles are needed for building robust ML models that can differentiate among MCI subtypes with higher accuracy.

To evaluate the completeness and discordance of outside rectal MRI initial staging reports compared to second-opinion reviews, and to assess the potential clinical impact of major discordance on treatment decisions in patients with rectal adenocarcinoma. A retrospective analysis of outside rectal MRI reviews submitted for second-opinion interpretation by subspecialized radiologists from June 2014–March 2020 was conducted. Outside and second review reports were compared side-by-side; cases with discordance (and those with major discordance, i.e., may alter treatment, particularly) were identified. Two colorectal surgeons, blinded to report origins, reviewed cases with major discordance to evaluate their theoretical impact on patient management and rated their confidence level of the reports on a five-point Likert scale (1=lowest confidence). In 461 patients (median age, 57 years [IQR: 49–67]; 274 male), compared to outside reviews, second reviews demonstrated improved report completeness across tumor characteristics, local extent, and nodal/metastatic disease clinical staging categories. The largest reporting gaps were in tumor morphology (66.4

Digital neuropsychological assessment easily captures behavior previously not obtainable by traditional pencil-and-paper tests. Verbal serial list learning tests are commonly used to assess for putative neurogenerative syndromes. Recognition test performance is often expressed compiling simple ‘yes/ no’ responses, but fail to assess process metrics such as the latency to respond to individual recognition test items. Memory clinic patients were assessed with the a digital neuropsychological protocol where cluster analysis of traditional metrics classified patients into normal (nl= 23), amnestic MCI (aMCI= 17), dysexecutive MCI (dMCI= 23), and dementia (dem= 14) groups. Verbal episodic memory was assessed with a 6-word Philadelphia (repeatable) Verbal Learning Test. P(r)VLT delayed recognition memory was assessed using a forced multiple-choice format where the iPad both displayed and verbally administered six trials containing the target word, a prototypic semantic foil (e.g., “apple”) and a generic semantic foil (e.g., “pear”). The patient was asked to touch the word that was part of the original word list as quickly as possible. aMCI and dementia groups endorsed more recognition foils that other groups (Table 1; dementia > all groups, p< 0.001; aMCI > dMCI & nl groups, p< 0.001). Serial list learning recognition latency was slower for aMCI versus dMCI and nl groups; p< 0.005); and dementia versus dMCI and nl groups (p< 0.001). Regression analysis (dv= recognition latency; block 1= age, education, sex; block 2= recognition prototypic & generic foils) was significant (R 2 = 0.463, p< 0.001); and found slower recognition latency was associated with greater numbers of prototypic recognition foils (beta= 0.551; p< 0.001). Regression analysis (dv= recognition latency; block 1= age, education, sex; block 2= all free recall cluster responses, all free recall extra-list intrusion errors) also found slower recognition latency was associated with fewer semantic cluster responses (beta= -0.252, p< 0.021), but greater numbers of extra-list intrusion errors (beta= 0.266, p< 0.017). When brought to scale, automated analysis of recognition latency along with other serial list learning process metrics could help identify early, emergent neurodegenerative illness.

Semantic memory refers to knowledge of attributes associated with common objects. Quantifying the strength of semantic association between successive ‘animal’ fluency responses can be challenging. The current research assessed between-group differences for ‘animal’ fluency total output and selected verbal serial list learning, episodic memory measures. Memory clinic patients were assessed with a digital neuropsychological protocol. Cluster analysis classified patients into normal (nl= 23), amnestic MCI (aMCI= 17), dysexecutive MCI (dMCI= 23), and dementia (dementia= 14) groups. During the protocol, patients were given 60secs to provide animal exemplars. Memory was assessed with the P(r)VLT, a 6-word verbal serial list learning test. Using artificial intelligence assisted scoring all ‘animal’ fluency and P(r)VLT outcome variables including the ‘animal’ Association Index (AI), where the mean number of shared attributes between successive responses were automatically tallied. The nl group generated more animal exemplars than all other groups (p< 0.001; Table 1). aMCI and dMCI patients generated more responses than dementia patients (p< 0.043, both analyses). NL and dMCI patients produced a higher, more semantically connected, ‘animal’ AI than dementia patients (nl > dem; p< 0.025, both analyses). Regression analysis (dv= ’animal’ AI; block 1= age, education, sex; block 2= recognition prototypic & generic foils) was significant (R 2 = 0.145, p< 0.023); and found that a reduced, more impaired ‘animal’ AI was associated with increasing numbers of P(r)VLT prototypic recognition foils (beta= -0.276; p< 0.024). Regression analysis (dv= ’animal’ AI; block 1= age, education, sex; block 2= P(r)VLT semantic cluster responses, P(r)VLT extra-list intrusion errors) was not significant. In addition to commonly used outcome measures such as total ‘animal’ responses, this digital neuropsychological protocol scores a number of process variables, including the ‘animal’ AI. The association between reduced ‘animal’ AI and greater numbers of list learning prototypic recognition foils suggests combined episodic memory and semantic-related impairment in selected patients. When brought to scale, automated analysis of neuropsychological process variables may aide in identifying emergent neurodegenerative illness.

BackgroundDual task paradigms are thought to offer a quantitative means to assess cognitive reserve and the brain’s capacity to allocate resources in the face of competing cognitive demands. The most common dual task paradigms examine the interplay between gait or balance control and cognitive function. However, gait and balance tasks can be physically challenging for older adults and may pose a risk of falls. ObjectiveWe introduce a novel, digital dual-task assessment that combines a motor-control task (the “ball balancing” test), which challenges an individual to maintain a virtual ball within a designated zone, with a concurrent cognitive task (the backward digit span task [BDST]). MethodsThe task was administered on a touchscreen tablet, performance was measured using the inertial sensors embedded in the tablet, conducted under both single- and dual-task conditions. The clinical use of the task was evaluated on a sample of 375 older adult participants (n=210 female; aged 73.0, SD 6.5 years). ResultsAll older adults, including those with mild cognitive impairment (MCI) and Alzheimer disease–related dementia (ADRD), and those with poor balance and gait problems due to diabetes, osteoarthritis, peripheral neuropathy, and other causes, were able to complete the task comfortably and safely while seated. As expected, task performance significantly decreased under dual task conditions compared to single task conditions. We show that performance was significantly associated with cognitive impairment; significant differences were found among healthy participants, those with MCI, and those with ADRD. Task results were significantly associated with functional impairment, independent of diagnosis, degree of cognitive impairment (as indicated by the Mini Mental State Examination [MMSE] score), and age. Finally, we found that cognitive status could be classified with >70% accuracy using a range of classifier models trained on 3 different cognitive function outcome variables (consensus clinical judgment, Rey Auditory Verbal Learning Test [RAVLT], and MMSE). ConclusionsOur results suggest that the dual task ball balancing test could be used as a digital cognitive assessment of cognitive reserve. The portability, simplicity, and intuitiveness of the task suggest that it may be suitable for unsupervised home assessment of cognitive function.

Background Distinguishing between mild cognitive impairment (MCI) and early dementia requires both neuropsychological and functional assessment that often relies on caregivers’ insights. Contacting a patient's caregiver can be time-consuming in a physician's already-filled workday. Objective To assess the utility of a brief, machine learning (ML)-enabled digital cognitive assessment, the Digital Clock and Recall (DCR), for detecting functional dependence. Methods We evaluated whether the DCR can help identify individuals at risk of functional deficits as measured by the informant-rated Functional Activities Questionnaire (FAQ) in older individuals including cognitively unimpaired, MCI, and dementia likely due to Alzheimer's disease. Results The DCR scaled well with FAQ scores, and ML classifiers trained on multimodal DCR features demonstrated strong performance in predicting functional impairment on a held-out test set. Differences in FAQ scores between DCR-predicted classes were comparable across key demographic groups. Conclusions The DCR can streamline the clinical decision-making, triage, and intervention planning associated with functional impairment in primary care.

INTRODUCTIONEarly detection of Alzheimer's disease and cognitive impairment is critical to improving the healthcare trajectories of aging adults, enabling early intervention and potential prevention of decline.METHODSTo evaluate multi-modal feature sets for assessing memory and cognitive impairment, feature selection and subsequent logistic regressions were used to identify the most salient features in classifying Rey Auditory Verbal Learning Test-determined memory impairment.RESULTSMultimodal models incorporating graphomotor, memory, and speech and voice features provided the stronger classification performance (area under the curve = 0.83; sensitivity = 0.81, specificity = 0.80). Multimodal models were superior to all other single modality and demographics models.DISCUSSIONThe current research contributes to the prevailing multimodal profile of those with cognitive impairment, suggesting that it is associated with slower speech with a particular effect on the duration, frequency, and percentage of pauses compared to normal healthy speech.

By 2050, 1 in 4 people worldwide will be living with hearing impairment. We propose a digital Speech Hearing Screener (dSHS) using short nonsense word recognition to measure speech-hearing ability. The importance of hearing screening is increasing due to the anticipated increase in individuals with hearing impairment globally. We compare dSHS outcomes with standardized pure-tone averages (PTA) and speech-recognition thresholds (SRT). Fifty participants (aged 55 or older underwent pure-tone and speech-recognition thresholding. One-way ANOVA was used to compare differences between hearing impaired and hearing not-impaired groups, by the dSHS, with a clinical threshold of moderately impaired hearing at 35 dB and severe hearing impairment at 50 dB. dSHS results significantly correlated with PTAs/SRTs. ANOVA results revealed the dSHS was significantly different (F(1,47) = 38.1, p < 0.001) between hearing impaired and unimpaired groups. Classification analysis using a 35 dB threshold, yielded accuracy of 85.7% for PTA-based impairment and 81.6% for SRT-based impairment. At a 50 dB threshold, dSHS classification accuracy was 79.6% for PTA-based impairment (Negative Predictive Value (NPV)-93%) and 83.7% (NPV-100%) for SRT-based impairment. The dSHS successfully differentiates between hearing-impaired and unimpaired individuals in under 3 min. This hearing screener offers a time-saving, in-clinic hearing screening to streamline the triage of those with likely hearing impairment to the appropriate follow-up assessment, thereby improving the quality of services. Future work will investigate the ability of the dSHS to help rule out hearing impairment as a cause or confounder in clinical and research applications.

Background By 2050, 1 in 4 people worldwide will be living with hearing impairment by 2050. We propose a digital Speech Hearing Screener (dSHS) using short nonsense word recognition to measure speech-hearing ability. We compare dSHS outcomes with standardized pure-tone averages (PTA) and speech-recognition thresholds (SRT). 50 participants (aged 55 or older underwent pure-tone and speech-recognition thresholding. Methods One-way ANOVA was used to compare differences between hearing impaired and hearing not-impaired groups, by the dSHS, with a clinical threshold of moderately impaired hearing at 35dB and severe hearing impairment at 50dB. Results dSHS results significantly correlated with PTAs/SRTs. ANOVA results revealed the dSHS was significantly different (F(1,47) = 38.1, p < 0.001) between hearing impaired and unimpaired groups. Classification analysis using a 35dB threshold, yielded accuracy of 85.7% forPTA-based impairment and 81.6% forSRT-based impairment. At a 50dB threshold, dSHS classification accuracy was 79.6% for PTA-based impairment (NPV-93%) and 83.7% (NPV-100%) for SRT-based impairment. Conclusions The dSHS successfully differentiates between hearing impaired and unimpaired individuals in under 3 minutes. This hearing screener offers a time saving, in clinic hearing screening to streamline the triage of those with likely hearing impairment to the appropriate follow up assessment, thereby improving the quality of services. Additionally, this tool can help to rule out hearing impairment as a cause or confounder of cognitive impairment.

Abstract Primary care providers currently wait for a complaint to initiate a ‘for-cause’ cognitive evaluation and often refer the patient to a specialist. Specialist evaluation then leads to the diagnosis, but treatment occurs late and is unlikely to meaningfully prevent or reduce disability. We hypothesized a brief digital cognitive assessment (DCA), the Digital Clock and Recall (DCR), could concurrently identify CI and predict PET Aβ status, thereby providing an efficient means for timely identification and prioritization of patients for disease-modifying treatments. 930 participants (age 72.0±6.7; 56.8% female; 23% minorities) were classified as cognitively unimpaired, mild cognitive impairment, or probable Alzheimer’s dementia, and 35.1% were Aβ+ on 18F-florbetapir PET scan. DCR-based models were compared with blood-based biomarkers (BBMs) Aβ42/40, pTau-181, and pTau-217, which poorly classified CI (AUCs=0.63, 0.66, 0.72) but accurately classified Aβ status (AUCs=0.81, 0.78, 0.89). DCR accurately classified CI and predicted Aβ status (AUCs=0.85, 0.83). Moreover, combining the DCR with Aβ42/40, pTau-181, and pTau-217 improved their Aβ-status prediction performance to AUCs of 0.882, 0.835, and 0.905, respectively. DCR was superior to BBMs in detecting cognitive impairment and boosted their Aβ-PET prediction ability to levels comparable to CSF biomarkers. We propose a workflow in which implementation of DCAs such as the DCR facilitates early identification of individuals with MCI or early dementia likely due to AD, thanks to the concurrent classification of CI and accurate prediction of Aβ PET status.

IntroductionAlzheimer’s disease and related dementias (ADRD) represent a substantial global public health challenge with multifaceted impacts on individuals, families, and healthcare systems. Brief cognitive screening tools such as the Mini-Cog© can help improve recognition of ADRD in clinical practice, but widespread adoption continues to lag. We compared the Digital Clock and Recall (DCR), a next-generation process-driven adaptation of the Mini-Cog, with the original paper-and-pencil version in a well-characterized clinical trial sample.MethodsDCR was administered to 828 participants in the Bio-Hermes-001 clinical trial (age median ± SD = 72 ± 6.7, IQR = 11; 58% female) independently classified as cognitively unimpaired (n = 364) or as having mild cognitive impairment (MCI, n = 274) or dementia likely due to AD (DLAD, n = 190). MCI and DLAD cohorts were combined into a single impaired group for analysis. Two experienced neuropsychologists rated verbal recall accuracy and digitally drawn clocks using the original Mini-Cog scoring rules. Inter-rater reliability of Mini-Cog scores was computed for a subset of the data (n = 508) and concordance between Mini-Cog rule-based and DCR scoring was calculated.ResultsInter-rater reliability of Mini-Cog scoring was good to excellent, but Rater 2’s scores were significantly higher than Rater 1’s due to variation in clock scores (p < 0.0001). Mini-Cog and DCR scores were significantly correlated (τB = 0.71, p < 0.0001). However, using a Mini-Cog cut score of 4, the DCR identified more cases of cognitive impairment (n = 47; χ2 = 13.26, p < 0.0005) and Mini-Cog missed significantly more cases of cognitive impairment (n = 87). In addition, the DCR correctly classified significantly more cognitively impaired cases missed by the Mini-Cog (n = 44) than vice versa (n = 4; χ2 = 21.69, p < 0.0001).DiscussionOur findings demonstrate higher sensitivity of the DCR, an automated, process-driven, and process-based digital adaptation of the Mini-Cog. Digital metrics capture clock drawing dynamics and increase detection of diagnosed cognitive impairment in a clinical trial cohort of older individuals.

Background Disease-modifying treatments for Alzheimer’s disease highlight the need for early detection of cognitive decline. However, at present, most primary care providers do not perform routine cognitive testing, in part due to a lack of access to practical cognitive assessments, as well as time and resources to administer and interpret the tests. Brief and sensitive digital cognitive assessments, such as the Digital Clock and Recall (DCR™), have the potential to address this need. Here, we examine the advantages of DCR over the Mini-Mental State Examination (MMSE) in detecting mild cognitive impairment (MCI) and mild dementia. Methods We studied 706 participants from the multisite Bio-Hermes study (age mean ± SD = 71.5 ± 6.7; 58.9% female; years of education mean ± SD = 15.4 ± 2.7; primary language English), classified as cognitively unimpaired (CU; n = 360), mild cognitive impairment (MCI; n = 234), or probable mild Alzheimer’s dementia (pAD; n = 111) based on a review of medical history with selected cognitive and imaging tests. We evaluated cognitive classifications (MCI and early dementia) based on the DCR and the MMSE against cohorts based on the results of the Rey Auditory Verbal Learning Test (RAVLT), the Trail Making Test-Part B (TMT-B), and the Functional Activities Questionnaire (FAQ). We also compared the influence of demographic variables such as race (White vs. Non-White), ethnicity (Hispanic vs. Non-Hispanic), and level of education (≥ 15 years vs. < 15 years) on the DCR and MMSE scores. Results The DCR was superior on average to the MMSE in classifying mild cognitive impairment and early dementia, AUC = 0.70 for the DCR vs. 0.63 for the MMSE. DCR administration was also significantly faster (completed in less than 3 min regardless of cognitive status and age). Among 104 individuals who were labeled as “cognitively unimpaired” by the MMSE (score ≥ 28) but actually had verbal memory impairment as confirmed by the RAVLT, the DCR identified 84 (80.7%) as impaired. Moreover, the DCR score was significantly less biased by ethnicity than the MMSE, with no significant difference in the DCR score between Hispanic and non-Hispanic individuals. Conclusions DCR outperforms the MMSE in detecting and classifying cognitive impairment—in a fraction of the time—while being not influenced by a patient’s ethnicity. The results support the utility of DCR as a sensitive and efficient cognitive assessment in primary care settings. Trial registration ClinicalTrials.gov identifier NCT04733989.

The prevalence of Alzheimer’s disease (AD) and related dementias (ADRD) is increasing. African Americans are twice as likely to develop dementia than other ethnic populations. Traditional cognitive screening solutions lack the sensitivity to independently identify individuals at risk for cognitive decline. The DCTclock is a 3-min AI-enabled adaptation of the well-established clock drawing test. The DCTclock can estimate dementia risk for both general cognitive impairment and the presence of AD pathology. Here we performed a retrospective analysis to assess the performance of the DCTclock to estimate future conversion to ADRD in African American participants from the Rush Alzheimer’s Disease Research Center Minority Aging Research Study (MARS) and African American Clinical Core (AACORE). We assessed baseline DCTclock scores in 646 participants (baseline median age = 78.0 ± 6.4, median years of education = 14.0 ± 3.2, 78% female) and found significantly lower baseline DCTclock scores in those who received a dementia diagnosis within 3 years. We also found that 16.4% of participants with a baseline DCTclock score less than 60 were significantly more likely to develop dementia in 5 years vs. those with the highest DCTclock scores (75–100). This research demonstrates the DCTclock’s ability to estimate the 5-year risk of developing dementia in an African American population. Early detection of elevated dementia risk using the DCTclock could provide patients, caregivers, and clinicians opportunities to plan and intervene early to improve cognitive health trajectories. Early detection of dementia risk can also enhance participant selection in clinical trials while reducing screening costs.

Radiologic imaging, especially MRI, has long been the mainstay for rectal cancer staging and patient selection for neoadjuvant therapy prior to surgical resection. In contrast, colonoscopy and CT have been the standard for colon cancer diagnosis and metastasis staging with T and N staging often performed at the time of surgical resection. With recent clinical trials exploring the expansion of the use of neoadjuvant therapy beyond the anorectum to the remainder of the colon, the current and future state of colon cancer treatment is evolving with a renewed interest in evaluating the role radiology may play in the primary T staging of colon cancer. The performance of CT, CT colonography, MRI, and FDG PET-CT for colon cancer staging will be reviewed. N staging will also be briefly discussed. It is expected that accurate radiologic T staging will significantly impact future clinical decisions regarding the neoadjuvant versus surgical management of colon cancer.