AI Group Research Articles

Autonomous Artificial Intelligence vs Artificial Intelligence–Assisted Human Optical Diagnosis of Colorectal Polyps: A Randomized Controlled Trial

BackgroundArtificial intelligence (AI)-based optical diagnosis systems (CADx) have been developed to allow pathology prediction of colorectal polyps during colonoscopies. However, CADx systems have not yet been validated for autonomous performance. Therefore, we conducted a trial comparing Autonomous AI to AI assisted human (AI-H) optical diagnosis. MethodsWe performed a randomized non-inferiority trial of patients undergoing elective colonoscopies in one academic institution. Patients were randomized int: 1) Autonomous AI-based CADx optical diagnosis of diminutive polyps without human input; 2) endoscopists performed optical diagnosis of diminutive polyps after seeing the real-time CADx diagnosis. Primary outcome was accuracy in optical diagnosis in both arms using pathology as gold standard. Secondary outcomes included agreement with pathology for surveillance intervals. Results467 patients were randomized (238 patients/158 polyps in the Autonomous AI group; 229 patients/179 polyps in the AI-H group). Accuracy for optical diagnosis was 77.2% (95%Confidence Interval 69.7-84.7) in the Autonomous AI group and 72.1% (95%CI 65.5-78.6) in the AI-H group (p=0.86). For high confidence diagnoses, accuracy for optical diagnosis was 77.2% (95%CI 69.7-84.7) in the Autonomous AI group and 75.5% (95%CI 67.9-82.0) in the AI-H group. Autonomous AI had statistically significantly higher agreement with pathology-based surveillance intervals compared to AI-H (91.5% [95%CI 86.9-96.1] vs 82.1% [95%CI 76.5-87.7]; p=0.016). ConclusionsAutonomous AI-based optical diagnosis exhibits non-inferior accuracy to endoscopist-based diagnosis. Both Autonomous AI and AI-H exhibited relatively low accuracy for optical diagnosis, however Autonomous AI achieved higher agreement with pathology-based surveillance intervals.

A Comparative Study of Responses to Retina Questions from Either Experts, Expert-Edited Large Language Models, or Expert-Edited Large Language Models Alone

ObjectiveTo assess the quality, empathy, and safety of expert edited large language model (LLM), human expert created and LLM responses to common retina patient questions DesignRandomized, masked multicenter study ParticipantsTwenty-one common retina patient questions were randomly assigned among 13 retina specialists. Each expert created a response (Expert) and then edited a LLM (ChatGPT-4)-generated response to that question (Expert+AI), timing themselves for both tasks. Five LLMs (ChatGPT-3.5, ChatGPT-4, Claude 2, Bing, Bard) also generated responses to each question. The original question along with anonymized and randomized Expert+AI, Expert and LLM responses were evaluated by the other experts who did not write an expert response to the question. Evaluators judged quality and empathy (very poor, poor, acceptable, good, or very good) along with safety metrics (incorrect information, likelihood to cause harm, extent of harm, and missing content). Main OutcomeMean quality and empathy score, proportion of responses with incorrect information, likelihood to cause harm, extent of harm, and missing content for each response type ResultsThere were 4008 total grades collected (2608 for quality and empathy; 1400 for safety metrics), with significant differences in both quality and empathy (p<0.001, p<0.001) between LLM, Expert and Expert+AI groups. For quality, Expert+AI (3.86 +/- 0.85) performed the best overall while GPT-3.5 (3.75 +/- 0.79) was the top performing LLM. For empathy, GPT-3.5 (3.75 +/- 0.69) had the highest mean score followed by Expert+AI (3.73 +/- 0.63). By mean score, Expert placed fourth out of seven for quality and sixth out of seven for empathy. For both quality (p<0.001) and empathy (p<0.001), expert-edited LLM responses performed better than expert-created responses. There were time savings for an expert-edited LLM response vs. expert-created response (p=0.02). ChatGPT-4 performed similar to Expert for Inappropriate Content (p=0.35), Missing Content (p=0.001), Extent of Possible Harm (p=0.356), and Likelihood of Possible Harm (p=0.129). Conclusions and RelevanceIn this randomized, masked, multicenter study, LLM responses were comparable to experts in terms of quality, empathy and safety metrics, warranting further exploration of their potential benefits in clinical settings.

Imaging study of coccygeal morphology in adolescent idiopathic scoliosis

The correlation between scoliosis and sagittal curvature of the cervical, thoracic, and lumbar spine have already been reported in previous studies. However, as a part of the spine, the change in coccygeal morphology in AIS patients has not yet been studied. In this study, a retrospective analysis was performed on 400 patients who were divided into a non-scoliotic group (206 patients) and an AIS group (194 patients). The Postacchini coccygeal radiological classification that was modified by Nathan was used to observe and compare the sagittal coccygeal morphology between the two groups. The results showed that the non-scoliotic group had the highest percentage (52.4%) of patients with type I and the lowest (3.4%) proportion of patients with type V; moreover, the AIS group had the highest percentage (69.1%) of patients with type I and the lowest (1.5%) proportion of patients with type V. The coccygeal morphology was significantly different between the non-scoliotic group and the AIS group (P = 0.001). No significant differences in coccygeal morphology were found between the males and females in the two groups (mild and moderate scoliosis and different segmental scoliosis). In addition, a significant correlation between coccygeal morphology and scoliosis (P = 0.035) was found. In conclusion, coccygeal morphology significantly differs between AIS patients and non-scoliotic adolescents. There was a smaller proportion of patients with a type I coccyx and a larger proportion of patients with a type II or type III coccyx in the AIS group than in the non-scoliotic group. In other words, the presence of a more pronounced coccygeal curve in AIS patients may be caused by an incorrect sitting position and an imbalance in the contraction of the pelvic muscles. It should be further studied whether correcting the sitting position and muscular imbalances could change coccygeal morphology and subsequently affect the development of AIS.

Could assisted reproductive techniques affect equine fetal membranes and neonatal outcome?

Embryo transfer (ET) and intracytoplasmic sperm injection (ICSI) are widely used in equine species, but their effects on fetal adnexa and neonates have not been investigated yet. The aim of this study was to retrospectively evaluate whether pregnancies obtained by ET or ICSI could be associated with the presence of macroscopic alterations of fetal membranes (FM) and umbilical cord (UC) and if the use of these techniques could influence neonatal outcome. Sixty-six light breed mares hospitalized at the Veterinary Teaching Hospital, University of Bologna, for attending delivery were included in the study. Mares were divided into Artificial Insemination (AI; 32/66 mares, 48 %), Embryo Transfer (ET; 12/66 mares, 18.2 %) and Intracytoplasmic Sperm Injection (ICSI; 22/66 mares, 33 %) groups. All the medical reports of mares and their foals were reviewed and data about mare, pregnancy, foaling, fetal membranes, umbilical cord and foal were recorded. The occurrence of dystocia resulted statistically different between AI group and ICSI group (p = 0.0066), and between AI group and ET group (p = 0.044). Macroscopic examination of FM revealed alterations in 30/66 mares (46 %): 8/32 in AI (25 %), 7/12 in ET (58 %) and 15/22 in ICSI (68 %) with significant lower incidence in AI compared to ET (p = 0.04) and ICSI (p = 0.002) groups. Alterations reported were chorionic villi hypoplasia, chorioallantois edema, allantois cysts, necrotic areas and greenish-grey concretions. Total length of UC resulted significantly shorter in ICSI group (49 ± 9 cm; p < 0.03) compared to AI (60 ± 17 cm) and ET (59 ± 15 cm). However, there were no differences in the incidence of foals' diseases at birth and in foals’ survival among groups (p > 0.05). The results demonstrate that transfer of in vivo or in vitro produced embryos may lead to alterations of placental development, as observed in other species, without being associated with a higher incidence of neonatal morbidity and mortality. Further studies about trophoblast development, FM histological evaluation, and placental gene expression should be carried out to clarify the mechanisms underlying the placental alterations.

Seasonal distribution of repeat breeder cows and evaluation of modified protocols for post AI treatment during summer.

An important problem of dairy industry worldwide is repeat breeder cows (RB). In this study, one farm in Northern Greece was chosen on the basis of subfertility. First, the available reproductive data of the previous year were evaluated; then, the farmer was advised to use the most common treatments for RB [GnRH with the third artificial insemination (AI+GnRH) or ovsynch in cows found empty after the second AI], and their efficacy was monitored throughout the year. Cows (n = 147) 2.5 to 4.5 years old were included. During the next year, post AI treatments (meta-ovsynch and meta-AI) were used during summer; GnRH (100 μg of gonadorelin) alone or in combination with meloxicam (0.5 mg/kg BW) was used during summer. Meta-ovsynch protocols were administered to 29 cows detected as non-pregnant after 3 AIs; cows were divided randomly into 3 meta-ovsynch (OS) groups: group OS (controls, n = 10), standard ovsynch protocol; group OS+GnRH (n = 10), OS plus GnRH 10 days after AI; and group OS+GnRH+NSAID (n = 9), OS plus GnRH 10 days after AI plus meloxicam 15 days after AI. Meta-AI protocols were tested in 27 RB cows; cows were divided randomly into 3 AI groups: GnRH (n = 9), 1 dose of GnRH with AI; GnRH-GnRH (n = 9), two doses of GnRH, one with AI and another 10 days later; and GnRH-GnRH-NSAID (n = 9), two doses of GnRH as in GnRH-GnRH group and one dose of meloxicam 15 days after AI. Pregnancy diagnosis was performed by ultrasound 38 days after AI. During the 1st year of evaluation, 53.79% of cows had received ≥ 3 AIs; the highest percentage (38.48%) of RB cows was recorded during autumn. After treatment for RB, the lowest (P < 0.05) pregnancy rate (PR) was recorded in August (18.18%) compared to January (75.0%), May (53.33%), and December (50.0%, P < 0.1). The efficacy of ovsynch during winter was significantly higher compared to all the other seasons. The percentage of cows with increased estrous cycle duration (24-37 days) after RB treatment and AI was the highest (P < 0.05) in spring and summer after AI+GnRH and the lowest (P < 0.05) during winter after ovsynch. Concerning the post AI treatments during summer, significantly higher pregnancy rate was recorded in the OS+GnRH+NSAID group (55.55%) compared to OS+GnRH (10%) and OS (10%) groups. Pregnancy rates did not differ among AI groups. It seems that OS+GnRH+NSAID is a promising protocol for RB cows during summer although further research is needed to support the present results.

Time efficiency, occlusal morphology, and internal fit of anatomic contour crowns designed by dental software powered by generative adversarial network: A comparative study

ObjectivesTo evaluate the time efficiency, occlusal morphology, and internal fit of dental crowns designed using generative adversarial network (GAN)-based dental software compared to conventional dental software. MethodsThirty datasets of partial arch scans for prepared posterior teeth were analyzed. Each crown was designed on each abutment using GAN-based software (AI) and conventional dental software (non-AI). The AI and non-AI groups were compared in terms of time efficiency by measuring the elapsed work time. The difference in the occlusal morphology of the crowns before and after design optimization and the internal fit of the crown to the prepared abutment were also evaluated by superimposition for each software. Data were analyzed using independent t tests or Mann–Whitney test with statistical significance (α=.05). ResultsThe working time was significantly less for the AI group than the non-AI group at T1, T5, and T6 (P≤.043). The working time with AI was significantly shorter at T1, T3, T5, and T6 for the intraoral scan (P≤.036). Only at T2 (P≤.001) did the cast scan show a significant difference between the two groups. The crowns in the AI group showed less deviation in occlusal morphology and significantly better internal fit to the abutment than those in the non-AI group (both P<.001). ConclusionsCrowns designed by AI software showed improved outcomes than that designed by non-AI software, in terms of time efficiency, difference in occlusal morphology, and internal fit. Clinical significanceThe GAN-based software showed better time efficiency and less deviation in occlusal morphology during the design process than the conventional software, suggesting a higher probability of optimized outcomes of crown design.

Anatomic trajectory for iliac screw placement adapts better to the morphological features of the pelvis of each individual than the S2 alar iliac screw: a radiological study

PurposeThe iliac fixation (IF) through the S2 ala permits the minimization of implant prominence and tissue dissection. An alternative to this technique is the anatomic iliac screw fixation (AI), which considers the perpendicular axis to the narrowest width of the ileum and the width of the screw. The morphological accuracy of the iliac screw insertion of two low profile iliac fixation (IF) techniques is investigated in this study.MethodsTwenty-nine patients operated on via low profile IF technique were divided into two groups, those treated using 28 screws with the starting point at S2, and those treated with 30 AI entry point. Radiological parameters (Tsv-angle, Sag-Angle, Max-length, sacral-distance, iliac-width, S2-midline, skin-distance, iliac-wing, and PSIS distance) and clinical outcomes (early and clinic complications) were evaluated by two blinded expert radiologists, and the results were compared in both groups with the real trajectory of the screws placed.ResultsDifferences between ideal and real trajectories were observed in 6 of the 9 evaluated parameters in the S2AI group. In the AI group, these trajectories were similar, except for TSV-Angle, Max-length, Iliac-width, and distance to iliac-wing parameters. Moreover, compared with S2AI, AI provided better adaptation to the pelvic morphology in all parameters, except for sagittal plane angulation, skin distance, and iliac width.ConclusionsAI ensures the advantages of low profile pelvic fixation like S2AI, with a starting point in line with S1 pedicle anchors and low implant prominence, and moreover adapts better to the morphological features of the pelvis of each individual.

The impact of in vitro embryo production on placental and umbilical cord vascularization is minimized by the addition of reproductive fluids

Animals born from in-vitro-produced (IVP) embryos show changes in the placenta and umbilical cord vascularization. This study compares the placental and umbilical vascular morphometry in pigs (n = 19) born through artificial insemination (AI group) or after transfer of IVP embryos cultured with (RF-IVP group) or without (C-IVP group) reproductive fluids. The relationship between vascular parameters and animal growth during the first year of life was also analyzed. Samples were collected at birth, fixed, paraffin-embedded, cut in sections, stained, and photographed for vascular and morphometric analysis with ImageJ® and Slide Viewer®. The average daily weight gain was individually scored from birth to the first year of life. No differences were found in placental vascular morphometry among groups, except for the vascular area of small vessels (arterioles, venules, and small vessels) that was higher in the C-IVP group. Regarding the umbilical cord, the values for perimeter (AI: 26.40 ± 3.93 mm; IVP: 30.51 ± 4.74 mm), diameter (AI: 8.35 ± 1.01 mm; IVP: 10.26 ± 1.85 mm), area (AI: 43.18 ± 12.87; IVP: 56.61 ± 14.89 mm2), and Wharton's jelly area (AI: 36.86 ± 12.04 mm2; IVP 48.88 ± 12.80 mm2) were higher in IVP-derived than AI-derived animals, whereas arterial and venous morphometric data were similar between groups. A correlation study showed that placental and umbilical cord vascular phenotypes affect the further growth of pigs. In conclusion, assisted reproductive technologies impact small caliber vessels in the placenta and morphometric parameters in the umbilical cord. The addition of reproductive fluids in IVP-embryo contributes to reduce the differences with in vivo-derived animals.

Comparative effectiveness of palbociclib plus aromatase inhibitor versus fulvestrant alone as initial endocrine therapy for HR+/HER2- advanced breast cancer in Chinese clinical practice: A real-world study.

e13057 Background: In the real world, there are still a large number of HR+/HER2- advanced breast cancer patients receiving initial endocrine therapy with fulvestrant monotherapy in China, especially patients with bone metastases only. It is due to lack of data comparing the effective between Palbociclib combined with AI and fulvestrant monotherapy as initial endocrine therapy. On the other hand, there are economic factors and access issues. According to the National Cancer Center's previous clinical experience, Palbociclib combined with AI is superior to fulvestrant monotherapy in HR+/HER2- advanced breast cancer patients receiving initial endocrine therapy. Therefore, we collected and summarized the previous clinical data in our center to compare the effective of Palbociclib combined with AI and fulvestrant monotherapy for HR+/ HER2-advanced breast cancer patients receiving initial endocrine therapy in China. Methods: This is a retrospective real-world experience aimed at verifying the effective of Palbociclib combined with AI and fulvestrant monotherapy for HR+/ HER2-advanced breast cancer patients receiving initial endocrine therapy in China. The primary aim was rwPFS; secondary aims were overall survival, and objective response rate. Patients received Palbociclib combined with AI or fulvsetrant monotherapy. Results: In total, 346 patients were enrolled, 193 to Palbociclib plus AI and 153 to fulvestrant monotherapy. Median age was 55 years (range 28-88); 191 had visceral involvement (116 in Palbociclib plus AI group and 75 in fulvestrant group). At a median follow up of 38 months (range 4-78). The median PFS was 20.0 months (95% CI 17.3–22.7) in the Palbociclib plus AI group and 12 months (95% CI 9.6–14.3) in the fulvestrant group (hazard ratio 0.593, 95% CI 0.470–0.749, p&lt;0·0001). The 6-month progression-free rate was 83.9% in the Palbociclib plus AI group and 75.8% in fulvestrant group. The 6-month progression-free rate was 83.9% in the Palbociclib plus AI group and 75.8% in fulvestrant group. The 12-month progression-free rate was 65.8% in the Palbociclib plus AI group and 46.4% in fulvestrant monotherapy group. The over survival was not reached in the Palbociclib plus AI group and it was 65 months (95% CI 55.6–75.4) in the fulvestrant group. Grade 3-4 adverse events occurred in Palbociclib plus AI group. The most common grade 3 or 4 adverse events were neutropenia (47.7%) and leucopenia (27.5%) in the Palbociclib plus AI group. Conclusions: The efficacy of palbociclib plus AI is worthy of recognition and the toxicity was acceptable. Compared with fulvestrant monotherapy in HR+/HER2- advanced breast cancer patients receiving initial endocrine therapy, Palbociclib plus AI might have a better survival. Clinical trial information: NCT05000736 .

Peri-procedural anesthesia and patient pain experience in pulmonary vein isolation by means of very high power short duration radiofrequency ablation

Abstract Funding Acknowledgements Type of funding sources: None. Background Very high-power short-duration (vHPSD) temperature-controlled radiofrequency ablation (vHPSD) may reduce ablation times and improve patient tolerability, permitting pulmonary vein (PV) isolation under mild conscious sedation. Purpose Evaluate the anesthetic drugs use and patients ‘pain experience during vHPSD PV isolation. Methods Thirty-three consecutive patients, with paroxysmal atrial fibrillation (AF), treated with QDot Micro catheter and vHPSD (90 w for 4 s) (vHPSD Group), were compared with the last 33 consecutive patients, with paroxysmal AF, treated with a surround flow contact force-sensing catheter guided by the ablation index (450 at anterior and superior sites at 50 W, 330 at posterior and inferior sites at 40 W) (AI Group). Anesthetic drugs (midazolam and fentanyl) use were compared as well as pain experience, measured using a 0-10 scale. Results All PVs were isolated. PV isolation ablation time in the vHPSD Group (5.9±1.4 min) was shorter than in the AI Group (25±11 min, p&amp;lt;0.001). Pain experience was significantly lower in vHPSD group (4.5±2 vs 6.6±1.8, p&amp;lt; 0.001). No patients required general anesthesia; fentanyl use was required in 4 vHPSD Group patients vs 25 AI Group patients (p&amp;lt;0.001). A lower dose of midazolam was required in vHPSD Group (2.3±1.2 mg vs 3.6±1.7 mg, p &amp;lt; 0.001). Conclusions vHPSD radiofrequency ablation for PVI can be performed under conscious sedation using only benzodiazepine in most of patients without compromising patient pain experience.

Approximate Unitary t-Designs by Short Random Quantum Circuits Using Nearest-Neighbor and Long-Range Gates

We prove that {{,textrm{poly},}}(t) cdot n^{1/D}-depth local random quantum circuits with two qudit nearest-neighbor gates on a D-dimensional lattice with n qudits are approximate t-designs in various measures. These include the “monomial” measure, meaning that the monomials of a random circuit from this family have expectation close to the value that would result from the Haar measure. Previously, the best bound was {{,textrm{poly},}}(t)cdot n due to Brandão–Harrow–Horodecki (Commun Math Phys 346(2):397–434, 2016) for D=1. We also improve the “scrambling” and “decoupling” bounds for spatially local random circuits due to Brown and Fawzi (Scrambling speed of random quantum circuits, 2012). One consequence of our result is that assuming the polynomial hierarchy ({{,mathrm{textsf{PH}},}}) is infinite and that certain counting problems are #{textsf{P}}-hard “on average”, sampling within total variation distance from these circuits is hard for classical computers. Previously, exact sampling from the outputs of even constant-depth quantum circuits was known to be hard for classical computers under these assumptions. However the standard strategy for extending this hardness result to approximate sampling requires the quantum circuits to have a property called “anti-concentration”, meaning roughly that the output has near-maximal entropy. Unitary 2-designs have the desired anti-concentration property. Our result improves the required depth for this level of anti-concentration from linear depth to a sub-linear value, depending on the geometry of the interactions. This is relevant to a recent experiment by the Google Quantum AI group to perform such a sampling task with 53 qubits on a two-dimensional lattice (Arute in Nature 574(7779):505–510, 2019; Boixo et al. in Nate Phys 14(6):595–600, 2018) (and related experiments by USTC), and confirms their conjecture that O(sqrt{n}) depth suffices for anti-concentration. The proof is based on a previous construction of t-designs by Brandão et al. (2016), an analysis of how approximate designs behave under composition, and an extension of the quasi-orthogonality of permutation operators developed by Brandão et al. (2016). Different versions of the approximate design condition correspond to different norms, and part of our contribution is to introduce the norm corresponding to anti-concentration and to establish equivalence between these various norms for low-depth circuits. For random circuits with long-range gates, we use different methods to show that anti-concentration happens at circuit size O(nln ^2 n) corresponding to depth O(ln ^3 n). We also show a lower bound of Omega (n ln n) for the size of such circuit in this case. We also prove that anti-concentration is possible in depth O(ln n ln ln n) (size O(n ln n ln ln n)) using a different model.

Quality of life in Spanish postmenopausal breast cancer patients with localized disease who finish endocrine treatment: a prospective study.

In this article, the quality of life (QOL) of Spanish postmenopausal early-stage breast cancer patients who have finished endocrine therapy (ET), QOL changes after endocrine therapy cessation, and the differences between two endocrine therapy modalities (tamoxifen or aromatase inhibitor [AI]) are studied. More QOL information after endocrine therapy cessation is needed. A prospective cohort study was performed. Participating in the study were 158 postmenopausal patients who had received tamoxifen or AI for 5 years. In some cases, endocrine therapy may have changed during those 5 years.Patients completed the European Organisation for Research and Treatment of Cancer QLQ-C30 and QLQ-BR45 questionnaires at baseline, after 6 months, and after 1 year of follow-up. Patients older than 65 years also completed the QLQ-ELD14. Linear mixed-effect models were used to evaluate longitudinal changes in QOL and differences in QOL between endocrine therapy modalities. QOL scores for the whole sample throughout follow-up were high (>80/100 points) in most QOL areas. Moderate limitations (>30 points) occurred in the QLQ-BR45 in sexual functioning and sexual enjoyment, future perspective, and joint symptoms. Moderate limitations also occurred in the QLQ-ELD14 in worries about others, maintaining purpose, joint stiffness, future worries, and family support. In those who had finished endocrine therapy, pain was reduced in all three assessments conducted during the 1-year follow-up period in both groups. Tamoxifen patients showed better QOL in functioning (role functioning, global QOL, financial impact), symptoms (pain), and emotional areas (future perspective and worries about others) than AI patients but worse QOL in skin mucosis symptoms. The results of this study show that postmenopausal early-stage breast cancer patients adapted well to their disease and endocrine therapy treatment. QOL improvements in the 1-year follow-up period appeared in one key area: pain. Differences between endocrine therapy modalities suggested QOL was better in the tamoxifen group than in the AI group.

Abstract P6-05-36: Cost-effectiveness of CDK4/6 inhibitors as a First line Therapy for Metastatic Breast Cancer. A Mexican Cohort

Abstract BACKGROUND. Breast cancer is the most frequent neoplasm worldwide, as reported by GLOBOCAN 2020, there were 2.2 million new cases per year and 680,000 deaths. In Mexico, it represents the leading cause of death from cancer in women, and therefore represents a public health problem in our country. The standard treatment for patients with hormone receptor-positive, her2-negative breast cancer is endocrine tehrapy with an aromatase inhibitor plus a CDK4/6 inhibitor (CDK4/6i+AI), however access to these therapies is difficult and limited resources in developing countries, lead to treatment strategies such as aromatase inhibitors alone (AI) or chemotherapy (ChT) still being used. However, management with ChT involves an increase in the use of reosurces due to cost per infusión, use of premedication and granulocyte colony-stimulating agents. OBJECTIVE The aim of this study was to provide an economic evaluation of CDK4/6i+AI compared with AI alone or ChT as a first line in MBC to better understand its value from the healthcare point of view in a developing country. METHODS. We designed a retrospective cost-effectiveness analysis of three different therapies CDK4/6i+AI, AI alone and ChT administered as first-line therapy for patients with MBC. RESULTS. A cost-effectiveness analysis was performed on a retrospective cohort of 150 MBC patients (march 2011 to April 2020) with a follow-up of al least 2 years. The median age at diagnosis was 55 years old. The utilization of health care resources was retrieved from clinical charts. Only direct costs associated with pre-progression, progression, and management of adverse events were considered and expressed on current USD. Seventy-six percent were diagnosed with de novo stage IV disease, 66% were postmenopausal and 76% had ductal histology. The most common sites of metastasis were visceral 55% and 29% had only bone metastases. We identified 3 treatment groups: (1) CDK4/6i+AI, 18.66% (28/150), (2) AI, 48.66% (73/150) and (3) ChT, 32.66% (49/150). The median PFS of iCDK4/6 + TH was 32.10 months compared with 18.87 (95%CI: 16.4, 28.7) months for the AI group and 6.57 months for chemotherapy. The HR of iCDK4/6+TH vs HT was 0.357 (95%CI: 0.18-0.72) and that of iCDK4/6+TH vs chemotherapy was 0.09 (95%CI: 0.04-0.22). Median OS survival was not reached in any arm. The most frequent adverse events grade 3 were fatigue 10.71%, neutropenia 32.14%, diarrhea 7.14%, myalgias 3.57% and arthralgias 3.57% in the CDK4/6i +AI group, fatigue 2.74% and arthralgias 4.11% in AI group and fatigue 20.41%, neutropenia 18.37%, nausea 10.2%, diarrhea 6.12%, myalgias 2.4% and headache 2.4% with chemotherapy. PFS was used as the outcome for the cost-effectiveness analysis, with 5 years of follow-up, CDK4/6i+AI offer an incremental efficacy of 1.4 years in PFS compared with AI and 2.43 years with ChT, they are related to an incremental cost of $28,151.61 and $26,720.47 concerning AI and ChT, respectively. The ICER for CDK4/6i+AI compared to AI is $20,108.29 and $10,996.07 compared to chemotherapy. CONCLUSION. CDK4/6i+AI increase years of life gained when compared to AI and chemotherapy. Is a cost-effective tratment in our institution because it is less than two GDP per capita. CDK4/6i+AI is the standar treatment around the world even in develop countries like Mexico. PFS 3 arms PFS 3 arms Citation Format: Maritza Ramos-Ramírez, Silvia Guzman-vazquez, Vanessa Dominguez-Esquivel, Jose Rodrigo Espinosa-Fernandez, Sandy Ruiz-Cruz, Paula Cabrera-Galeana, Alexandra Garcilazo, Luis Antonio Cabrera-Miranda, Claudia Haydee Arce Salinas. Cost-effectiveness of CDK4/6 inhibitors as a First line Therapy for Metastatic Breast Cancer. A Mexican Cohort. [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P6-05-36.

A Preliminary Study on Microbiota Characteristics of Bronchoalveolar Lavage Fluid in Patients with Pulmonary Nodules Based on Metagenomic Next-Generation Sequencing

The characteristics and roles of microbes in the occurrence and development of pulmonary nodules are still unclear. We retrospectively analyzed the microbial mNGS results of BALF from 229 patients with pulmonary nodules before surgery, and performed a comparative analysis of lung flora between lung cancer and benign nodules according to postoperative pathology. The analysis also focused on investigating the characteristics of lung microbiota in lung adenocarcinomas with varying histopathology. There were differences in lung microbiota between lung cancer and benign lung nodules. Bacterial diversity was lower in lung cancer than in benign lung nodules. Four species (Porphyromonas somerae, Corynebacterium accolens, Burkholderia cenocepacia and Streptococcus mitis) were enriched in lung cancer compared with the benign lung nodules. The areas under the ROC curves of these four species were all greater than 0.6, and the AUC of Streptococcus mitis was 0.702, which had the highest diagnostic value for differentiating lung cancer from benign lung diseases. The significantly enriched microbiota varied with the different pathological subtypes of lung adenocarcinoma. Streptococcus mitis, Burkholderia oklahomensis and Burkholderia latens displayed a trend of increasing from the benign lung disease group to the AIS group, MIA group and IAC group, whereas Lactobacillus plantarum showed a downward trend. Changes in the abundance of lung microbiota are closely related to the development of infiltrating adenocarcinoma. Our findings provide new insights into the relationship between the changes in lung microbiota and the development of lung cancer.

The effectiveness of real-time computer-aided and quality control systems in colorectal adenoma and polyp detection during colonoscopies: a meta-analysis.

This study was conducted adhering to PRISMA guidelines. Studies were searched across PubMed, CINAHL, EMBASE, Scopus, Cochrane, and Web of Science. Keywords including the following 'Artificial Intelligence, Polyp, Adenoma, Detection, Rate, Colonoscopy, Colorectal, Colon, Rectal' were used. Odds ratio (OR) applying 95% CI for PDR and ADR were computed. SMD with 95% CI for withdrawal times were computed using RevMan 5.4.1 (Cochrane). The risk of bias was assessed using the RoB 2 tool. Of 2562 studies identified, 11 trials were included comprising 6856 participants. Of these, 57.4% participants were in the AI group and 42.6% individuals were in in the standard group. ADR was higher in the AI group compared to the standard of care group (OR=1.51, P=0.003). PDR favored the intervened group compared to the standard group (OR=1.89, P<0.0001). A medium measure of effect was found for withdrawal times (SMD=0.25, P<0.0001), therefore with limited practical applications. AI-supported colonoscopies improve PDR and ADR; however, no noticeable worsening of withdrawal times is noted. Colorectal cancers are highly preventable if diagnosed early-on. With AI-assisted tools in clinical practice, there is a strong potential to reduce the incidence rates of cancers in the near future.

Breakthrough of error correction in quantum computing

By using a superconducting processor, Google Quantum AI group demonstrated that a logic qubit realized by surface code of quantum error correction performs better when the number of physical qubits increases. Two surface codes to encode logic qubits for scaling are realized experimentally, with multiple cycles of quantum error correcting assisted by ancillary qubits. This result can be considered as an important step toward fault-tolerant quantum computers. In this review paper, we introduce briefly the mechanism of quantum error correction. As an example, Shor’s nine-qubit error correction code is explained. Then, the new experiments of Google quantum AI group are introduced to show their significance in scaling. The advances in other quantum error correction experiments are also reviewed. Finally, the development of quantum computers is discussed.

Artificial intelligence assisted discrimination between pulmonary tuberculous nodules and solid lung cancer nodules

The differential diagnosis between pulmonary tuberculous nodules and solid lung cancer nodules is difficult and easy to be misdiagnosed in clinic. The data of clinic and image features of Chest CT with 70 cases of non-calcified pulmonary tuberculous nodules and 198 cases of solid lung cancer nodules confirmed by pathology in the Department of Thoracic Surgery or Respiratory and Critical Care Medicine, the First Affiliated Hospital of Chongqing Medical University from January to September 2020 were collected retrospectively. The characteristics of clinical and chest CT were compared between pulmonary tuberculous nodules and solid lung cancer nodules. The sensitivity, specificity, accuracy and negative predictive value in the two groups were compared between Artificial Intelligence assisted diagnosis system and manual image reading. The results found that the mean age, past tumor history, family history of tumor, CT image features of nodules includes mean diameter, short burr, blood vessel crossing in the pulmonary tuberculous nodules group were lower than those in the solid lung cancer group (p < 0.05). In 35 cases of pulmonary tuberculous nodules group and 63 cases of solid lung cancer nodules group with Dicom format thin-slice chest CT, the sensitivity of AI-assisted diagnosis was 98.98 %. The diagnosis specificity, accuracy and negative predictive value in the AI group (80.61 %, 92.06 %, 60.00 %) were much higher than these in the intermediate respiratory physicians (62.24 %, 76.19 %, 37.14 %, p = 0.004, 0.015, 0.044) respectively, and there was no significant difference between AI and senior radiologists. There are many similarities in clinical and CT image features between pulmonary tuberculous nodules and solid lung cancer nodules. The ability of AI-assisted diagnosis system is better than that of intermediate physicians, reaching the diagnostic level of senior physicians, which is conducive to homogenization and improvement of the differential diagnosis ability of physicians.

Minority social influence and moral decision-making in human–AI interaction: The effects of identity and specialization cues

In group decision-making, the behavior of each member is sensitive to the social influence of other majority members. Research on majority influence has shown that multiple non-human agents with anthropomorphic cues can exert normative pressure on a lone human decision-maker. However, how individuals perceive and respond to minority influence exerted by a lone machine is rarely discussed. Hence, a between-subjects experiment was conducted to examine how different minority identity (human vs artificial intelligence [AI]) and specialization (specialist vs generalist) cues influence individuals’ perceptions and behavior in response to moral dilemmas in a joint human–AI group. The results confirmed the significant role of specialization cues in predicting in-group identification, source credibility, and conversion behavior. In addition, the participants perceived the human minority as more credible than the AI minority, which prompted conversion behavior when the minority was labeled as a specialist rather than as a generalist.

The Clinical Value of Systemic Inflammatory Response Index and Inflammatory Prognosis Index in Predicting 3-Month Outcome in Acute Ischemic Stroke Patients with Intravenous Thrombolysis.

PurposeSystemic inflammatory response index (SIRI) was an independent predictor of the prognosis of many diseases. Inflammatory prognostic index (IPI) was a new inflammatory prognostic marker with certain clinical significance. We aimed to explore the association between SIRI, IPI and early stage severity of stroke as well as 3-month outcome of AIS patients.Patients and MethodsA total of 63 AIS patients who treated with alteplase were selected. The patients were divided into mild group and moderate to severe group according to the National Institutes of Health Stroke Scale (NIHSS) scores. According to the modified Rankin scale (mRS) score, patients were divided into the good prognosis group and the poor prognosis group. Spearman correlation statistically analyzed the correlation between SIRI, IPI and NIHSS score. Univariate and multivariate logistic regression analyzed the risk factors of 3-month prognosis. ROC curve was adopted to predict the effect of SIRI and IPI levels on poor prognosis in AIS patients.ResultsSpearman analysis showed that there was positively correlated with NIHSS score and IPI in mild AIS group (r=0.541, P<0.05). Compared with the mild group, SIRI and IPI in the moderate to severe group was significantly higher (P<0.05). The SIRI and IPI in the poor prognosis group were significantly higher than those in the good prognosis group (P<0.05). Univariate and multivariate logistic regression analysis showed that SIRI and IPI were independent prognostic factors for the 3-month prognosis of AIS patients (P< 0.05). The ROC curve showed that the areas under the 3-month prognosis curve predicted by SIRI and IPI were 0.685, 0.774 respectively.ConclusionIPI is correlated with stroke severity at admission. SIRI and IPI are independent predictors of short-term prognosis in AIS patients. SIRI and IPI can be a novel the good short-term prognostic biomarker for AIS patients treated with intravenous thrombolysis.

P410 Aspergillus fumigatus sensitization among the patients with chronic obstructive pulmonary disease (COPD)— a cross sectional study

Poster session 3, September 23, 2022, 12:30 PM - 1:30 PMObjectivesAllergic bronchopulmonary aspergillosis (ABPA) caused by hypersensitivity to A. fumigatus complicates the course of asthma. Fungal sensitization due to A. fumigatus among asthmatic patients and their progression to ABPA is well studied. Similar data on Aspergillus sensitization among patients with COPD and their progression are still not well established. The objective of this study was to evaluate the Total serum IgE (TIgE) and A. fumigatus specific IgE (Af sp IgE) levels among patients with COPD.MethodA total of 100 stable patients with COPD above 40 years of age from the Department of Pulmonary Medicine were included. TIgE and Af sp IgE levels were detected using VIDAS total IgE assay and M3 ImmunoCap with Phadia 100 respectively. The subjects were grouped into three (TIgE <500 IU/l, 500-1000 IU/l, and >1000 IU/l) based on the TIgE values. They were also categorized based on Af sp IgE levels as sensitization likely (AS ≥0.35 kUA/l), sensitization indeterminate (AI 0.1-0.35 kUA/l), and sensitization unlikely (AU ≤0.1 kUA/l). This categorization was based on the kit manufacturer's guidelines that sensitization is unlikely with specific IgE <0.1 kUA/l and the proposed ISHAM ABPA working group criteria of >0.35 kUA/l for diagnosis of sensitization. A comparison of Asf sp IgE with TIgE was done using the Fischer exact test.ResultAmong 100 patients, the prevalence of elevated TIgE [>150 IU/l (kit cutoff)] and Aspergillus sensitization [Af Sp IgE >0.35kUA/l (ISHAM cutoff)] was 47% and 6%, respectively. A total of 5% of subjects satisfied the criteria for serological ABPA. The results of the comparison of Asf sp IgE with total IgE using the Fischer exact test are given in Table 1.AS category: TIgE was >1000 IU/l in 5 (83.3%) of the 6 subjects. One (16.7%) had between 500-1000 IU/l. None of them had TIgE <500IU/l. Aspergillus sensitization was higher in 3/5 subjects with TIgE >1000 IU/ml (Asf sp IgE values of 2.55KUA/l, 12KUA/l, 14KUA/l, 0.38KUA/l, and 0.45KUA/l) compared with the one subject with TIgE <1000IU/l (Asf sp IgE of 0.73KUA/l). The clinical characteristics of subjects in the AS category is given in Table 2.AI category: A total of 7 (50%) of the 14 subjects had the TIgE value >1000 IU/l, 5 (35.7%) had the TIgE between 500-1000 IU/l, and 2 subjects had the TIgE <500 IU/l.AU category: In all, 1 (1.25%) had the TIgE value >1000 IU/l, 5 (6.25%) had the TIgE between 500-1000 IU/l. TIgE was <500 IU/l in 74 (92.5%) subjects.ConclusionThe study results suggest the co-existence of COPD and Aspergillus sensitization/ABPA. Patients in the AI group (Af sp IgE level 0.1-0.35 kUA/l) must be evaluated and monitored to prevent the progression of the disease. Studies involving a larger patient population are warranted.