An AI-based approach driven by genotypes and phenotypes to uplift the diagnostic yield of genetic diseases

Identifying disease-causing variants in Rare Disease patients' genome is a challenging problem. To accomplish this task, we describe a machine learning framework, that we called "Suggested Diagnosis", whose aim is to prioritize genetic variants in an exome/genome based on the probability of being disease-causing. To do so, our method leverages standard guidelines for germline variant interpretation as defined by the American College of Human Genomics (ACMG) and the Association for Molecular Pathology (AMP), inheritance information, phenotypic similarity, and variant quality. Starting from (1) the VCF file containing proband's variants, (2) the list of proband's phenotypes encoded in Human Phenotype Ontology terms, and optionally (3) the information about family members (if available), the "Suggested Diagnosis" ranks all the variants according to their machine learning prediction. This method significantly reduces the number of variants that need to be evaluated by geneticists by pinpointing causative variants in the very first positions of the prioritized list. Most importantly, our approach proved to be among the top performers within the CAGI6 Rare Genome Project Challenge, where it was able to rank the true causative variant among the first positions and, uniquely among all the challenge participants, increased the diagnostic yield of 12.5% by solving 2 undiagnosed cases.

Perivascular Fat Density and Contrast Plaque Enhancement: Does a Correlation Exist?

BACKGROUND AND PURPOSE

Inflammatory changes in the fat tissue surrounding the coronary arteries have been associated with coronary artery disease and high-risk vulnerable plaques. Our aim was to investigate possible correlations between the presence and degree of perivascular fat density and a marker of vulnerable carotid plaque, namely contrast plaque enhancement on CTA.

MATERIALS AND METHODS

One-hundred patients (76 men, 24 women; mean age, 69 years) who underwent CT angiography for investigation of carotid artery stenosis were retrospectively analyzed. Contrast plaque enhancement and perivascular fat density were measured in 100 carotid arteries, and values were stratified according to symptomatic (ipsilateral-to-cerebrovascular symptoms)/asymptomatic status (carotid artery with the most severe degree of stenosis). Correlation coefficients (Pearson ρ product moment) were calculated between the contrast plaque enhancement and perivascular fat density. The differences among the correlation ρ values were calculated using the Fisher r-to-z transformation. Mann-Whitney analysis was also calculated to test differences between the groups.

RESULTS

There was a statistically significant positive correlation between contrast plaque enhancement and perivascular fat density (ρ value = 0.6582, P value = .001). The correlation was stronger for symptomatic rather than asymptomatic patients (ρ value = 0.7052, P value = .001 versus ρ value = 0.4092, P value = .001).

CONCLUSIONS

There was a positive association between perivascular fat density and contrast plaque enhancement on CTA. This correlation was stronger for symptomatic rather than asymptomatic patients. Our results suggest that perivascular fat density could be used as an indirect marker of plaque instability.

The role of a commercial radiation dose index monitoring system in establishing local dose reference levels for fluoroscopically guided invasive cardiac procedures

PURPOSE

The primary goal was to evaluate local dose level for fluoroscopically guided invasive cardiac procedures in a high-volume activity catheterization laboratory, using automatic data registration with minimal impact on operator workload. The secondary goal was to highlight the relationship between dose indices and acquisition parameters, in order to establish an effective strategy for protocols optimization.

METHODS

From September 2016 to December 2018, a dosimetric survey was conducted in the 2 rooms of the catheterization laboratory of our institution. Data collection burden was minimized using a commercial Radiation Dose Index Monitoring System (RDIMs) that analyzes dicom files automatically sent by the x-ray equipment. Data were combined with clinical information extracted from the HIS records reported by the interventional cardiologist. Local dose levels were established for different invasive cardiac procedures.

RESULTS

A total of 3029 procedures performed for 2615 patients were analyzed. Median KAP were 21 Gycm² for invasive coronary angiography (ICA) procedures, 61 Gycm² for percutaneous coronary intervention (PCI) procedures, 59 Gycm² for combined (ICA+PCI) procedures, 87 Gycm² for structural heart intervention (TAVI) procedures. A significant dose reduction (51% for ICA procedures and 58% for PCI procedures) was observed when noise reduction acquisition techniques were applied.

CONCLUSIONS

RDIMs are effective tools in the establishment of local dose level in interventional cardiology, as they mitigate the burden to collect and register extensive dosimetric data and exposure parameters. Systematic review of data support the multi-disciplinary team in the definition of an effective strategy for protocol management and dose optimization.

A 512-channels, whole array readout, CMOS implantable probe for acute recordings from the brain

The integration of implantable CMOS neural probes with thousands of simultaneously recording microelectrodes is a promising approach for neuroscience and might allow to literally image electrophysiological neuronal activity in multiple brain circuits as we have previously shown in vitro. Here, we present a complete system based on a fully multiplexed CMOS neural probe that was designed for in-vivo acute recordings with a scalable circuit architecture. In particular, a first prototype of a single-shaft probe with 512 electrodes was realized in a standard CMOS 0.18μm technology and post-processed to structure the shaft with a wedge-like geometry of 30μm in thickness at the tip and 80μm at the base. The design of the system and of the probe as well as the post-processing techniques are discussed. Finally, preliminary results on electrical, mechanical and implantation tests are presented to demonstrate the feasibility of our approach.

Step-and-Shoot IMRT by Siemens Beams: An EPID Dosimetry Verification During Treatment

PURPOSE

This work reports the extension of a semiempirical method based on the correlation ratios to convert electronic portal imaging devices transit signals into in vivo doses for the step-and-shoot intensity-modulated radiotherapy Siemens beams. The dose reconstructed at the isocenter point Diso, compared to the planned dose, Diso,TPS, and a γ-analysis between 2-dimensional electronic portal imaging device images obtained day to day, seems to supply a practical method to verify the beam delivery reproducibility.

METHOD

The electronic portal imaging device images were obtained by the superposition of many segment fields, and the algorithm for the Diso reconstruction for intensity-modulated radiotherapy step and shoot was formulated using a set of simulated intensity-modulated radiotherapy beams. Moreover, the in vivo dose-dedicated software was integrated with the record and verify system of the centers.

RESULTS

Three radiotherapy centers applied the in vivo dose procedure at 30 clinical intensity-modulated radiotherapy treatments, each one obtained with 5 or 7 beams, and planned for patients undergoing radiotherapy for prostatic tumors. Each treatment beam was checked 5 times, obtaining 900 tests of the ratios R = Diso/Diso,TPS. The average R value was equal to 1.002 ± 0.056 (2 standard deviation), while the mean R value for each patient was well within 5%, once the causes of errors were removed. The γ-analysis of the electronic portal imaging device images, with 3% 3 mm acceptance criteria, showed 90% of the tests with Pγ < 1 ≥ 95% and γmean ≤ 0.5. The off-tolerance tests were found due to incorrect setup or presence of morphological changes. This preliminary experience shows the great utility of obtaining the in vivo dose results in quasi real time and close to the linac, where the radiotherapy staff may immediately spot possible causes of errors. The in vivo dose procedure presented here is one of the objectives of a project, for the development of practical in vivo dose procedures, financially supported by the Istituto Nazionale di Fisica Nucleare.

Free-breathing conformal irradiation of pancreatic cancer

The purpose of this study was to assess treatment margins in free-breathing irradiation of pancreatic cancer after bone alignment, and evaluate their impact on conformal radiotherapy. Fifteen patients with adenocarcinoma of the head of the pancreas underwent implantation of single fiducial marker. Intrafraction uncertainties were assessed on simulation four-dimensional computed tomography (4D CT) by calculating maximal intrafraction fiducial excursion (MIFE). In the first ten patients, after bony alignment, the position of the fiducial was identified on weekly acquired megavolt cone-beam CT (MV-CBCT). The interfraction residual uncertainties were estimated by measuring the fiducial displacements with respect to the position in the first session. Patient mean (pM) and patient standard deviation (pSD) of fiducial displacement, mean (μM) and standard deviation (μSD) of pM, and root-mean-square of pSD (σ(res)) were calculated. In the other five patients, MIFE was added to the residual component to obtain personalized margin. In these patients, conformal kidney sparing (CONKISS) irradiation was planned prescribing 54/45 Gy to PTV1/PTV2. The organ-at-risk limits were set according to current NCCN recommendation. No morbidity related to the fiducial marker implantation was recorded. In the first ten patients, along right-left, anterior-posterior, and inferior-superior directions, MIFE was variable (mean ± std = 0.24 ± 0.13 cm, 0.31 ± 0.14 cm, 0.83 ± 0.35 cm, respectively) and was at most 0.51, 0.53, and 1.56 cm, respectively. Along the same directions, μM were 0.09, -0.05, -0.05 cm, μSD were 0.30, 0.17, 0.33 cm, and σ(res) were 0.35, 0.26, and 0.30 cm, respectively. MIFE was not correlated with pM and pSD. In the five additional patients, it was possible to satisfy recommended dose limits, with the exception of slightly higher doses to small bowel. After bony alignment, the margins for target expansion can be obtained by adding personalized MIFE to the residual interfraction term. Using these margins, conformal free-breathing irradiation is a reliable option for the treatment of pancreatic cancer.

Quasi-VMAT in high-grade glioma radiation therapy

PURPOSE

To compare a quasi-volumetric modulated arc therapy (qVMAT) with three-dimensional conformal radiation therapy (3D-CRT) and intensity-modulated radiation therapy (IMRT) for the treatment of high-grade gliomas. The qVMAT technique is a fast method of radiation therapy in which multiple equispaced beams analogous to those in rotation therapy are radiated in succession.

PATIENTS AND METHODS

This study included 12 patients with a planning target volume (PTV) that overlapped at least one organ at risk (OAR). 3D-CRT was planned using 2-3 non-coplanar beams, whereby the field-in-field technique (FIF) was used to divide each field into 1-3 subfields to shield the OAR. The qVMAT strategy was planned with 15 equispaced beams and IMRT was planned using 9 beams with a total of 80 segments. Inverse planning for qVMAT and IMRT was performed by direct machine parameter optimization (DMPO) to deliver a homogenous dose distribution of 60 Gy within the PTV and simultaneously limit the dose received by the OARs to the recommended values. Finally, the effect of introducing a maximum dose objective (max. dose < 54 Gy) for a virtual OAR in the form of a 0.5 cm ring around the PTV was investigated.

RESULTS

The qVMAT method gave rise to significantly improved PTV95% and conformity index (CI) values in comparison to 3D-CRT (PTV95% = 90.7 % vs. 82.0 %; CI = 0.79 vs. 0.74, respectively). A further improvement was achieved by IMRT (PTV95% = 94.4 %, CI = 0.78). In qVMAT and IMRT, the addition of a 0.5 cm ring around the PTV produced a significant increase in CI (0.87 and 0.88, respectively), but dosage homogeneity within the PTV was considerably reduced (PTV95% = 88.5 % and 92.3 %, respectively). The time required for qVMAT dose delivery was similar to that required using 3D-CRT.

CONCLUSION

These findings suggest that qVMAT should be preferred to 3D-CRT for the treatment of high-grade gliomas. The qVMAT method could be applied in hospitals, for example, which have limited departmental resources and are not equipped with systems capable of VMAT delivery.

Prostate image-guided radiotherapy by megavolt cone-beam CT

PURPOSE

To test megavolt cone-beam CT (MV-CBCT) in order to evaluate setup errors in prostate radiotherapy.

PATIENTS AND METHODS

The setup of 9 patients was verified weekly by electronic portal imaging (EPI) and MV-CBCT, both per-formed in the same treatment session. EPI were compared with digitally reconstructed radiographies (DRRs). MV-CBCTs were matched to simulation CTs by manual registration based on bone markers (BMR), by manual registration based on soft tissues (STR) - rectum, bladder, and seminal vesicles - and by automatic registration (AR) performed by a mutual information algorithm. Shifts were evaluated along the three main axes: anteroposterior (AP), craniocaudal (CC), and laterolateral (LL). Finally, in 4 additional patients showing intraprostatic calcifications, the calcification mismatch error was used to evaluate the three MV-CBCT matching methods.

RESULTS

A total of 50 pairs of orthogonal EPIs and 50 MV-CBCTs were analyzed. Assuming an overall tolerance of 2 mm, no significant differences were observed comparing EPI vs BMR in any axis. A significant difference (p < 0.001) was observed along the AP axis comparing EPI vs AR and EPI vs STR. On the calcification data set (22 measures), the calcification mismatch along the AP direction was significantly lower (p < 0.05) after STR than after BMR or AR.

CONCLUSION

Bone markers were not an effective surrogate of the target position and significant differences were observed comparing EPI or BMR vs STR, supporting the assessment of soft tissue position by MVCBs to verify and correct patient setup in prostate radiotherapy.

Application of a practical method for the isocenter pointin vivodosimetry by a transit signal

This work reports the results of the application of a practical method to determine the in vivo dose at the isocenter point, D(iso), of brain thorax and pelvic treatments using a transit signal S(t). The use of a stable detector for the measurement of the signal S(t) (obtained by the x-ray beam transmitted through the patient) reduces many of the disadvantages associated with the use of solid-state detectors positioned on the patient as their periodic recalibration, and their positioning is time consuming. The method makes use of a set of correlation functions, obtained by the ratio between S(t) and the mid-plane dose value, D(m), in standard water-equivalent phantoms, both determined along the beam central axis. The in vivo measurement of D(iso) required the determination of the water-equivalent thickness of the patient along the beam central axis by the treatment planning system that uses the electron densities supplied by calibrated Hounsfield numbers of the computed tomography scanner. This way it is, therefore, possible to compare D(iso) with the stated doses, D(iso,TPS), generally used by the treatment planning system for the determination of the monitor units. The method was applied in five Italian centers that used beams of 6 MV, 10 MV, 15 MV x-rays and (60)Co gamma-rays. In particular, in four centers small ion-chambers were positioned below the patient and used for the S(t) measurement. In only one center, the S(t) signals were obtained directly by the central pixels of an EPID (electronic portal imaging device) equipped with commercial software that enabled its use as a stable detector. In the four centers where an ion-chamber was positioned on the EPID, 60 pelvic treatments were followed for two fields, an anterior-posterior or a posterior-anterior irradiation and a lateral-lateral irradiation. Moreover, ten brain tumors were checked for a lateral-lateral irradiation, and five lung tumors carried out with three irradiations with different gantry angles were followed. One center used the EPID as a detector for the S(t) measurement and five pelvic treatments with six fields (many with oblique incidence) were followed. These last results are reported together with those obtained in the same center during a pilot study on ten pelvic treatments carried out by four orthogonal fields. The tolerance/action levels for every radiotherapy fraction were 4% and 5% for the brain (symmetric inhomogeneities) and thorax/pelvic (asymmetric inhomogeneities) irradiations, respectively. This way the variations between the total measured and prescribed doses at the isocenter point in five fractions were well within 2% for the brain treatment, and 4% for thorax/pelvic treatments. Only 4 out of 90 patients needed new replanning, 2 patients of which needed a new CT scan.

Decline of maternal hepatitis A virus antibody levels in infants

Hepatitis A is a common viral infection causing substantial morbidity and mortality. The anti-hepatitis A virus (HAV) vaccination in infants would guarantee control of the infection. However, the immunogenicity of the HAV vaccine in infants could be impaired by the presence of passively acquired maternal HAV antibodies. This study evaluated the prevalence of HAV antibodies in 103 women at delivery and in their babies in the first year of life. Eighteen mothers (17.5%) had anti-HAV serum level >10 mIU ml(-1). In their infants the anti-HAV level was still positive in 11 out of 18 (61.1%) at 12 mo. Two out of 85 infants born to anti-HAV-negative mothers and anti-HAV negative at birth were found to be positive at 5 mo of age.

CONCLUSION

It is proposed that all women be screened at delivery for anti-HAV antibodies. Children born to anti-HAV-negative mothers could be vaccinated early during the first year of life, whereas vaccination could be postponed in children born to anti-HAV-positive mothers, if necessary.

Prospective multicenter study on the clinical utility of ca-72.4 in postmenopausal patients with pelvic mass

The study objective was to evaluate the sensitivity and specificity as well as the positive predictive value and negative predictive value of CA 72.4 and CA 125 determination, separately and in combination, for diagnosing ovarian tumors in post-menopausal women with pelvic mass. The 299 patients recruited in this study underwent gynecological examination, plasma determination of CA 72.4 and CA 125, and laparotomy with histological definition of pelvic mass. CA 72.4 assay values were under 3.9 U/ml in 194 cases (70.8%); values ranged from 3.9 to 4.5 U/ml in 7 cases (2.5%) and were greater than 4.5 U/ml in 73 cases (26.6%). CA 72.4 assay was positive (>4.5 U/ml) in 56 cases (57.1%) of malignant ovarian pathology, in 4 cases (25%) of malignant extra-ovarian pathology as well as in 9 cases (7.1%) of benign ovarian pathology and in 4 cases (11.8%) of benign extra-ovarian pathology. With a cut-off at 3.9 U/ml, CA 72.4 showed a specificity of 91.3% and a sensitivity of 62.2%, whereas with a cut-off at 4.5 U/ml specificity was 92.9% and sensitivity 57.1%. Results of CA 125 assay for diagnosing a pelvic neoplasia (ovarian or extra-ovarian), showed a specificity of 85.3% and sensitivity of 68.8%. The agreement of the two markers (CA 125 and CA 72.4) as negative or positive shows a specificity of 77% and a sensitivity of 84.7% for ovarian cancer and a specificity of 73.5% and sensitivity of 75% for the diagnosis of pelvic neoplasias.

Prospective multicenter study on CA 125 in postmenopausal pelvic masses

The purpose of this study was to determine the diagnostic value of CA 125 in comparison with transabdominal ultrasound (US) in the evaluation of postmenopausal women with pelvic mass to detect malignant epithelial ovarian tumors. Postmenopausal patients with pelvic mass were studied with gynecologic examination, US and CA 125 determination. Three hundred eighty-eight patients were entered in the study. According to stratification based on US (probably benign, equivocal, possibly malignant) and CA 125 (< 35 U/ml, negative; between 35 and 65 U/ml, borderline; > 65 U/ml, positive), 290 patients were considered eligible for surgery. Specificity, sensitivity, positive and negative predictive value, and accuracy of US and CA 125 were calculated with respect to histological examination. Out of 290 operated patients, 134 had a benign ovarian pathology, 34 had extraovarian benign pathology, 106 had an ovarian malignancy, and 16 presented with an extraovarian malignant pathology. The results according to ovarian malignant pathology were as follows. CA 125 (> 65 U/ml): Specificity, 92.5%; sensitivity, 71.7%; accuracy, 83.3%. CA 125 (> 35 U/ml): Specificity, 82.0%; sensitivity, 78.3%; accuracy, 80.4%. US: Specificity, 77.6%; sensitivity, 84.9%; accuracy, 80.3%. Combination of US and CA 125 (> 65 U/ml): Specificity, 96.1%; sensitivity, 91.7%; accuracy, 94.3%. Determination of CA 125 is a highly specific method in predicting ovarian cancer in postmenopausal women with a pelvic mass. The association with US significantly improves the overall accuracy and may support therapeutical decision making by distinguishing between a significant percentage of women most likely to benefit from prompt intervention and women who may be managed following minor surgical diagnostic approach, such as fine-needle aspiration.

Absent end-diastolic velocity in umbilical artery: risk of neonatal morbidity and brain damage

OBJECTIVE

We conducted a cohort study in growth-retarded fetuses to establish if absent or reverse end-diastolic flow in the umbilical artery was associated with increased perinatal mortality and morbidity and neurologic damage at long-term follow-up.

STUDY DESIGN

Thirty-one fetuses with intrauterine growth retardation and absent or reverse end-diastolic flow in the umbilical artery (study group) and 40 growth-retarded fetuses with detectable diastolic flow in the umbilical artery, divided into two control groups, were followed up with serial nonstress tests, Doppler flow studies, and biophysical profiles. Twenty newborns from the study group survived the perinatal period and were observed for a mean of 18 months (range 12 to 24 months). Their neurologic outcomes were compared with those of 26 neonates from the two control groups.

RESULTS

Study group fetuses had a higher incidence of abnormal karyotype (9.7% vs 0%) and corrected perinatal mortality (26% vs 6% and 4%) and a greater risk of permanent neurologic sequelae (35% vs 0% and 12%) compared with the fetuses from the two control groups.

CONCLUSIONS

Growth-retarded fetuses with absent or reverse end-diastolic flow in the umbilical artery not only have an increased fetal and neonatal mortality but also a higher incidence of long-term permanent neurologic damage when compared with growth-retarded fetuses with diastolic flow in the umbilical circulation.

The role of sonographically guided aspiration in the clinical management of ovarian cysts

This study was undertaken to verify the role of fine-needle aspiration (FNA) followed by cytologic examination as a possible alternative to surgery in case of cystic pelvic masses. From January 1988 to March 1989, 204 patients with a proven cystic pelvic mass underwent FNA under sonographic guidance. In 20 cases the aspiration was performed transvaginally. Thirty-six patients were postmenopausal. In all cases the aspirated fluid was collected for cytologic evaluation. Other than one case of persistent hematuria, no complications occurred. The overall recurrence rate, verified 3 months from FNA, was 65%. Fifty-two per cent of patients developed a new cyst after a complete aspiration. Fifty-three patients underwent a laparotomy, allowing a comparison between cytologic and histologic patterns. The sensitivity of cytologic examination of aspirated fluids was 40% (3 of 5 malignancies were missed) and the specificity was 100% (no false positives were observed). We conclude that FNA might be proposed in young women with a unilocular ovarian cyst to avoid a surgical procedure. In postmenopausal women with a unilocular cystic mass 5 cm or less, FNA may be considered as an important step in the diagnosis. In all instances the ultrasonographic appearance of the cyst (echo texture and regularity of wall) and the characteristics of aspirated fluid are the most important findings. When the aspirated fluid contains mucus or blood, or when a complex mass is present, exploratory laparotomy or a laparoscopy is recommended.