Radiology report generation with medical knowledge and multilevel image-report alignment: A new method and its verification.

Medical report generation is an integral part of computer-aided diagnosis aimed at reducing the workload of radiologists and physicians and alerting them of misdiagnosis risks. In general, medical report generation is an image captioning task. Since medical reports have long sequences with data bias, the existing medical report generation models lack medical knowledge and ignore the interaction alignment between the two modalities of reports and images. The current paper attempts to mitigate these deficiencies by proposing an approach based on knowledge enhancement with multilevel alignment (MKMIA). To this end, it includes a knowledge enhancement (MKE) module and a multilevel alignment module (MIRA). Specifically, the MKE deals with general medical knowledge (MK) and historical knowledge (HK) obtained via data training. The general knowledge is embedded in the form of a dictionary with characteristic organs (referred to as Key) and organ aliases, disease symptoms, etc. (referred to as Value). It provides explicit exception candidates to mitigate data bias. Historical knowledge ensures the comparison of similar cases to provide a better diagnosis. MIRA furnishes coarse-to-fine multilevel alignment, reducing the gap between image and text features, improving the knowledge enhancement module's performance, and facilitating the generation of lengthy reports. Experimental results on two radiology report datasets (i.e., IU X-ray and MIMIC-CXR) proved the effectiveness of the proposed approach, achieving state-of-the-art performance.

Diagnostic value of the radiographic muscle-to-bone thickness ratio between the pronator quadratus and the distal radius at the same level in undisplaced distal forearm fracture.

PURPOSE: To investigate the value of the muscle-to-bone ratio (MBR), which is the ratio between the maximum pronator quadratus thickness (mPQT) and the distal radial thickness (DRT) at the same level, in comparison with mPQT in the diagnosis of undisplaced fractures in distal forearm (FDF) on radiographs. MATERIAL AND METHODS: This study was conducted with 210 wrists of 210 volunteers (79 male, 131 female) and 106 wrists of 106 patients with acute undisplaced FDF (53 male, 53 female). The mPQT was measured on lateral radiographs, and the MBR was calculated for each wrist. RESULTS: The mean mPQT (7.37 ± 1.99 mm) and MBR (0.55 ± 0.15) in the fracture group were significantly larger than those (4.70 ± 1.42 mm and 0.31 ± 0.09) in the control group respectively (all P<0.05). Sexual differences of mPQT were found in both control and fracture groups (all P<0.05). There were no gender differences of MBR in either control or fracture group (all P>0.05). ROC analysis revealed that the discriminating efficacy of MBR was significantly superior to that of mPQT (P<0.01). The most optimal diagnostic discrimination was obtained by a MBR cut-off value of 0.4. CONCLUSIONS: The MBR can be used as a new index for the diagnosis of undisplaced FDF, as well as a complementary means to detect occult fractures. If the MBR is larger than 0.4 in a patient with suspected or occult FDF, a further MRI examination will be justifiably needed.

Cochlear nerve diameters on multipoint measurements and effects of aging in normal-hearing children using 3.0-T magnetic resonance imaging.

OBJECTIVE: The aim of this study was to determine differences in diameter of cochlear nerves (CN) among three measurement points at the midpoint of the internal auditory canal (IAC), IAC fundus and cochlear aperture, and to evaluate whether nerve size varies with age. METHODS: A total of 336 normal-hearing ears of 201 children were assessed, who were underwent 3D-Fiesta sequence scanning of inner ear. All subjects were divided into 12 groups at one year interval. The diameter measurements of CN were obtained in the midpoint of the IAC, IAC fundus and cochlear aperture respectively on the axial and oblique sagittal images of 3.0-T MRI. SPSS 18.0 statistics software was applied for data analysis, and all of the data showed a normal distribution and expressed in x ± s. RESULTS: The diameters of normal-hearing children's CN at the midpoint of the IAC, IAC fundus and cochlear aperture were respectively: 1.12 ± 0.08 mm, 1.05 ± 0.06 mm, 0.87 ± 0.14 mm, and there were significant differences among the three measuring points (F = 527.57, p < 0.05). The diameters of the CN had no significant difference (p > 0.05) in age groups, gender and sides (p > 0.05), and there was no correlation between the diameters of normal children's CN and age (r is -0.129, 0.128 and -0.113, respectively). CONCLUSION: The diameters of normal-hearing children's CN change with different points of the internal auditory canal, of which the maximum value is in the midpoint of the IAC, followed by the IAC fundus, and the cochlear aperture is at the minimum; moreover the normal size doesn't change with age.

Role of scavenger receptors in dendritic cell function.

Dendritic cells (DCs), the most potent of the antigen-presenting cells, are crucial in initiating and shaping innate and adaptive immune responses. DCs discriminate unmodified self antigens from non-self and altered/modified self antigens via a large family of receptors called pattern-recognition receptors, which include Toll-like receptors and scavenger receptors (SRs). Recent findings underscore the critical role of SRs on DCs in pathogen clearance, atherosclerosis, apoptotic cell recognition, diesel exhaust particle recognition, etc. These new findings present SRs as an unexplored therapeutic target that warrants further basic and applied research, and have implications for vaccine development. This review highlights recent insights into the emerging role of these receptors in DC-mediated immune responses.

HRCT and MRI findings in X-linked non-syndromic deafness patients with a POU3F4 mutation.

OBJECTIVE: The aim of this study was to analyze HRCT and MRI findings in patients with X-linked non-syndromic deafness and a POU3f4 mutation. METHODS: HRCT and MRI data of four patients (males, 2-19 years old) with a POU3f4 mutation were collected and a retrospective review was performed. Cochlea, internal auditory canal (IAC), vestibule, semicircular canals, vestibular aqueduct, nerve canals in the IAC fundus, stapes and cochlear nerve were evaluated on 2D images (multi-planner reformation, MPR) and cochlear foramen on 3D images (CT virtual endoscopy, CTVE). Ten cases with normal hearing subjected to CT and MR exams served as controls. RESULTS: Inner ear malformations were bilateral and symmetrical. Cochlear malformation was shown to consist of as a relatively normal outer coat shape, absence of a cochlear modilous, and a direct intercommunication between the IAC and cochlear inner cavity. The lateral portion of the IAC was dilated. A spiral cochlear inner cavity was observed with CTVE images versus a helical cochlear nerve foramen as seen in controls. The labyrinthine facial nerve canal and superior vestibular nerve canal were enlarged. The Bill's bar was hypertrophic and partially pneumatized. A thickened stapes footplate was present and a fissura ante fenestram was absent in seven ears examined. A column shaped stapes was observed in one ear. CONCLUSIONS: The absence of a cochlear modilous with a dilated lateral IAC and thickened stapes footplate were the remarkable features observed with imaging these in X-linked non-syndromic deafness patients with a POU3F4 mutation. Preoperative recognition of the image features in these patients is important because it precludes stapedectomy and indicates the risks in the surgery of cochlear implantation including CSF gusher and electrode insertion into IAC.

Detection of hypoplasia of bony cochlear nerve canal by virtual endoscopy: a pilot study.

BACKGROUND: Dimensions of the bony cochlear nerve canal (BCNC) have been proposed as a potential diagnostic standard for hypoplasia of BCNC, but the standard remains inconsistent. We have previously found that a helix-like shape appears in normal BCNCs at VE images, whereas, the sign does not appear in some hypoplastic BCNCs. PURPOSE: To retrospectively examine the feasibility of computed tomographic (CT) virtual endoscopy (VE) in the evaluation of hypoplasia of BCNC on the basis of absence of a helix-like shape. MATERIAL AND METHODS: Twenty ears in 14 consecutive patients (mean age 5.5 years, range 1-15 years, 6 boys, 8 girls) diagnosed with hypoplasia of BCNC were included in this work. One hundred ears in 50 gender- and age-matched individuals (mean age 6.6 years, range 1-15 years, 29 boys, 21 girls) without inner ear disease and internal auditory canal (IAC) malformations served as controls. The presence or absence of a helix-like shape was evaluated by two independent reviewers. The value of VE for the diagnosis of hypoplasia of BCNC was assessed with clinical results and routine radiologic evaluation as the reference standard. Inter-observer agreement was calculated. Sensitivity, specificity, and accuracy were selected to test the diagnostic ability of the VE. RESULTS: Absence of a helix-like shape was found in the cochlear area of 17 of 20 ears in patients with hypoplasia of BCNC but in none of the control subjects. Inter-observer agreement was substantial (к = 0.773). The diagnostic rates of absence of a helix-like shape for hypoplasia of BCNC in terms of sensitivity, specificity, and accuracy were 85%, 100%, and 98%, respectively. There were significant differences between the two groups with respect to VE findings for absence of a helix-like shape (P < 0.001). CONCLUSION: The absence of a helix-like shape at VE images may be used as a potentially useful sign in the diagnosis of hypoplasia of BCNC.

[Observation of foramina hypoplasia within internal auditory canal fundus with CT virtual endoscopy].

OBJECTIVE: To observe normal and abnormal findings of foramina within internal auditory canal (IAC) fundus with Virtual Endoscopy of multisection helical Computed Tomography. METHOD: CT scanning of temporal bone was undergone in 25 volunteers (50 ears) in the normal group, in 8 cases (13 ears) with foramina hypoplasia within IAC fundus which were shown on CT and MRI images. CT virtual endoscopy (CTVE) findings were observed in the normal and abnormal groups. The lower threshold value was 900-1200 HU, the upper value was 3,071 HU in CTVE definitions. RESULT: The shape and position of foramina within IAC fundus were shown on the CTVE image, 50 foramina within IAC fundus were shown in all normal cases. In abnormal group, 9 ears were shown as IAC fundus disorder and foramina absent, only 1 foramina was shown in 1 ear, 2 foramen in 4 ears, 3 foramen in 2 ears, 4 foramen in 2 ears; Abnormal shape of cochlear nerve foramina was seen in 4 ears. Other malformations were also shown, including IAC malformations in 3 ears and inner ear malformations in 8 ears. CONCLUSION: CTVE is useful to show the shape and position of normal foramina and pathological changes in the patients with foramina hypoplasia within IAC fundus.