Finding the limits of deep learning clinical sensitivity with fractional anisotropy (FA) microstructure maps.

Background: Quantitative maps obtained with diffusion weighted (DW) imaging, such as fractional anisotropy (FA) -calculated by fitting the diffusion tensor (DT) model to the data,-are very useful to study neurological diseases. To fit this map accurately, acquisition times of the order of several minutes are needed because many noncollinear DW volumes must be acquired to reduce directional biases. Deep learning (DL) can be used to reduce acquisition times by reducing the number of DW volumes. We already developed a DL network named "one-minute FA," which uses 10 DW volumes to obtain FA maps, maintaining the same characteristics and clinical sensitivity of the FA maps calculated with the standard method using more volumes. Recent publications have indicated that it is possible to train DL networks and obtain FA maps even with 4 DW input volumes, far less than the minimum number of directions for the mathematical estimation of the DT. Methods: Here we investigated the impact of reducing the number of DW input volumes to 4 or 7, and evaluated the performance and clinical sensitivity of the corresponding DL networks trained to calculate FA, while comparing results also with those using our one-minute FA. Each network training was performed on the human connectome project open-access dataset that has a high resolution and many DW volumes, used to fit a ground truth FA. To evaluate the generalizability of each network, they were tested on two external clinical datasets, not seen during training, and acquired on different scanners with different protocols, as previously done. Results: Using 4 or 7 DW volumes, it was possible to train DL networks to obtain FA maps with the same range of values as ground truth - map, only when using HCP test data; pathological sensitivity was lost when tested using the external clinical datasets: indeed in both cases, no consistent differences were found between patient groups. On the contrary, our "one-minute FA" did not suffer from the same problem. Conclusion: When developing DL networks for reduced acquisition times, the ability to generalize and to generate quantitative biomarkers that provide clinical sensitivity must be addressed.

Structural Optimization Design of Microfluidic Chips Based on Fast Sequence Pair Algorithm.

The market for microfluidic chips is experiencing significant growth; however, their development is hindered by a complex design process and low efficiency. Enhancing microfluidic chips' design quality and efficiency has emerged as an integral approach to foster their advancement. Currently, the existing structural design schemes lack careful consideration regarding the impact of chip area, microchannel length, and the number of intersections on chip design. This inadequacy leads to redundant chip structures resulting from the separation of layout and wiring design. This study proposes a structural optimization method for microfluidic chips to address these issues utilizing a simulated annealing algorithm. The simulated annealing algorithm generates an initial solution in advance using the fast sequence pair algorithm. Subsequently, an improved simulated annealing algorithm is employed to obtain the optimal solution for the device layout. During the wiring stage, an advanced wiring method is used to designate the high wiring area, thereby increasing the success rate of microfluidic chip wiring. Furthermore, the connection between layout and routing is reinforced through an improved layout adjustment method, which reduces the length of microchannels and the number of intersections. Finally, the effectiveness of the structural optimization approach is validated through six sets of test cases, successfully achieving the objective of enhancing the design quality of microfluidic chips.

Identifying Susceptibility Loci for Cutaneous Squamous Cell Carcinoma Using a Fast Sequence Kernel Association Test.

Cutaneous squamous cell carcinoma (cSCC) accounts for about 20% of all skin cancers, the most common type of malignancy in the United States. Genome-wide association studies (GWAS) have successfully identified multiple genetic variants associated with the risk of cSCC. Most of these studies were single-locus-based, testing genetic variants one-at-a-time. In this article, we performed gene-based association tests to evaluate the joint effect of multiple variants, especially rare variants, on the risk of cSCC by using a fast sequence kernel association test (fastSKAT). The study included 1,710 cSCC cases and 24,304 cancer-free controls from the Nurses' Health Study, the Nurses' Health Study II and the Health Professionals Follow-up Study. We used UCSC Genome Browser to define gene units as candidate loci, and further evaluated the association between all variants within each gene unit and disease outcome. Four genes HP1BP3, DAG1, SEPT7P2, and SLFN12 were identified using Bonferroni adjusted significance level. Our study is complementary to the existing GWASs, and our findings may provide additional insights into the etiology of cSCC. Further studies are needed to validate these findings.

Use of fast-sequence spine MRI in pediatric patients.

OBJECTIVE: The immediate and long-term risk of anesthesia in the pediatric population is controversial. Traditional spine MRI protocols require the patient to remain still during the examination, and in young children this frequently results in the need for sedation administration. The authors' goal was to develop an abbreviated spine MRI protocol to reduce sedation administration in young patients undergoing spine MRI. METHODS: After IRB approval, the medical records of all pediatric patients who underwent a fast spine MRI protocol between 2017 and 2019 were reviewed. The protocol consisted of T2-weighted half-Fourier acquisition single-shot turbo spin echo, T1-weighted turbo spin echo, and T2-weighted STIR sequences acquired in the sagittal plane. The total acquisition time was 2 minutes with no single sequence acquisition longer than 60 seconds. Interpretability of the scans was assessed in accordance with the radiology report in conjunction with the neurosurgeon's clinical notes. RESULTS: A total of 47 fast spine MRI sessions were performed in 45 patients. The median age at the time of the MRI was 2.4 years (25th-75th quartile, 1.1-4.3 years; range 0.16-18.58 years). The most common indication for imaging was to rule out or follow a known syrinx (n = 30), followed by the need to rule out or follow known spinal dysraphism (n = 22). There were no uninterpretable or unusable scans. Eight of 47 scans were noted to have moderate motion artifact limitations with respect to the quality of the scan. Seven patients underwent a subsequent MRI with a sedated standard spine protocol within 1 year from the fast scan, which confirmed the findings on the fast MRI protocol with no new findings identified. CONCLUSIONS: The authors report the first pediatric series of a fast spine MRI protocol for use in young patients. The protocol does not require sedation and is able to identify and monitor syrinx, spinal dysraphism, and potentially other intraspinal anomalies.

Reliability of CSF turbulence and choroid plexus visualization on fast-sequence MRI in pediatric hydrocephalus.

OBJECTIVE Endoscopic third ventriculostomy with choroid plexus cauterization for the treatment of neonatal and infant hydrocephalus has gained popularity in the past decade. Identifying treatment failure is critically important. Results of a pilot study of 2 novel imaging markers seen on fast-sequence T2-weighted axial MRI showed potential clinical utility. However, the reliability of multiple raters detecting these markers must be established before a multicenter validation study can be performed. METHODS Two sets of de-identified single-shot T2-weighted turbo spin-echo axial images were prepared from scans of patients before and after they underwent endoscopic third ventriculostomy with choroid plexus cauterization between March 2013 and January 2016. The first set showed the lateral and third ventricles for visualization of turbulent CSF dynamics, and the second set showed the lateral ventricular atria for choroid plexus glomus detection. Three raters (Group 1) received written instructions before evaluating each image set once and then again 1 week later. Another 8 raters (Group 2) evaluated both image sets after oral instruction and group training on a pretest image set. Fleiss' kappa coefficients with 95% CIs were calculated for intrarater and interrater reliability in Group 1 and interrater reliability in Group 2. RESULTS Intrarater reliability kappa coefficients for Group 1 were ≥ 0.74 for turbulence and ≥ 0.80 for choroid plexus; their interrater kappa coefficients at the initial assessment were 0.50 (95% CI 0.37-0.62) and 0.56 (95% CI 0.43-0.69), respectively. The Group 2 interrater kappa scores were 0.82 (95% CI 0.78-0.86) for turbulence and 0.62 (95% CI 0.58-0.66) for choroid plexus. CONCLUSIONS With minimal training, intrarater reliability on visualization of turbulence and the choroid plexus was substantial, but interrater reliability was only moderate. After modestly increasing training, interrater reliability improved to near perfect and to substantial reliability for visualization of turbulence and choroid plexus, respectively. Given adequately trained observers, a multicenter investigation into the validity and potential clinical utility of the imaging markers seems feasible.

Children and Adults Both Learn Motor Sequences Quickly, But Do So Differently.

Both children and adults can learn motor sequences quickly in one learning session, yet little is known about potential age-related processes that underlie this fast sequence acquisition. Here, we examined the progressive performance changes in a one-session modified serial reaction time task in 6- and 10-year-old children and adults. We found that rapid sequence learning, as reflected by reaction time (RT), was comparable between groups. The learning was expressed through two behavioral processes: online progressive changes in RT while the task was performed in a continuous manner and offline changes in RT that emerged following a short rest. These offline and online RT changes were age-related; learning in 6-year-olds was primarily reflected through the offline process. In contrast, learning in adults was reflected through the online process; and both online and offline processes occurred concurrently in 10-year-olds. Our results suggest that early rapid sequence learning has a developmental profile. Although the unifying mechanism underlying these two age-related processes is unclear, we discuss possible explanations that need to be systematically elucidated in future studies.

Radiographic markers of clinical outcomes after endoscopic third ventriculostomy with choroid plexus cauterization: cerebrospinal fluid turbulence and choroid plexus visualization.

OBJECTIVE Endoscopic third ventriculostomy (ETV) with choroid plexus (CP) cauterization (CPC) represents a viable treatment option for congenital hydrocephalus in infants younger than 2 years. Imaging studies complement clinical data in the evaluation of treatment success or failure. The objectives of this study were to investigate novel radiographic markers-cerebrospinal fluid (CSF) turbulence and CP visualization-and their ability to reflect or predict clinical outcomes following ETV/CPC. METHODS Hydrocephalic patients younger than 2 years who were initially treated by ETV/CPC at the senior authors' institution between March 2013 and February 2014 were retrospectively reviewed. Clinical data, as well as the visualization of CSF turbulence and CP on pre- and postoperative fast-sequence MRI, were recorded. Radiographic images were reviewed by a blinded observer based on specific criteria for the visualization of CSF turbulence and CP. Data were collected and analyzed using descriptive statistics, including Fisher's exact test for comparisons. The research team obtained appropriate institutional review board approval for this study, without the need for informed consent. RESULTS Among the 32 patients (53% male and 47% female) studied, 18 of 32 (56%) responded favorably to initial or repeat ETV/CPC, with 13 of 32 (41%) patients requiring 1 surgery. Of the 19 (59%) patients whose initial ETV/CPC failed, 8 of 19 (42%) patients underwent repeat ETV/CPC, with 5 of 8 (63%) patients responding favorably. Radiographic CSF turbulence appeared more frequently following ETV/CPC failure than after ETV/CPC success (55% vs 18%, respectively; p = 0.02). The sensitivity and specificity of CSF turbulence as a radiographic marker for ETV/CPC failure were 80% and 58%, respectively. The radiographic depiction of CP disappearance following ETV/CPC from pre- to postoperative imaging occurred in 20 of 30 patients (67%). Among the patients who responded unsuccessfully to ETV/CPC and ultimately required secondary shunt insertion, 71% (10 of 14 patients) demonstrated CP persistence on postoperative imaging. In contrast, 6% (1 of 18) of patients who were treated successfully by ETV/CPC demonstrated the presence of CP on follow-up imaging. This difference reached statistical significance (p = 0.0001). The visualization of CP persistence despite ETV/CPC reflected treatment failure with 91% sensitivity and 81% specificity. The sensitivity of either or both radiographic markers to suggest ETV/CPC failure was 77%, while their specificity (both markers absent, thereby indicating ETV/CPC success) was 81%. CONCLUSIONS Radiographic markers correlate with clinical outcomes following the treatment of infantile hydrocephalus with ETV/CPC. Specifically, CSF turbulence may indicate ongoing pathological CSF flow dynamics, while CP absence following ETV/CPC may predict shunt independence. Future studies that incorporate prospective review and formal intra- and interobserver reliability estimates may help corroborate the utility of these radiographic markers.