Diabetic foot ulcers segmentation challenge report: Benchmark and analysis

Monitoring the healing progress of diabetic foot ulcers is a challenging process. Accurate segmentation of foot ulcers can help podiatrists to quantitatively measure the size of wound regions to assist prediction of healing status. The main challenge in this field is the lack of publicly available manual delineation, which can be time consuming and laborious. Recently, methods based on deep learning have shown excellent results in automatic segmentation of medical images, however, they require large-scale datasets for training, and there is limited consensus on which methods perform the best. The 2022 Diabetic Foot Ulcers segmentation challenge was held in conjunction with the 2022 International Conference on Medical Image Computing and Computer Assisted Intervention, which sought to address these issues and stimulate progress in this research domain. A training set of 2000 images exhibiting diabetic foot ulcers was released with corresponding segmentation ground truth masks. Of the 72 (approved) requests from 47 countries, 26 teams used this data to develop fully automated systems to predict the true segmentation masks on a test set of 2000 images, with the corresponding ground truth segmentation masks kept private. Predictions from participating teams were scored and ranked according to their average Dice similarity coefficient of the ground truth masks and prediction masks. The winning team achieved a Dice of 0.7287 for diabetic foot ulcer segmentation. This challenge has now entered a live leaderboard stage where it serves as a challenging benchmark for diabetic foot ulcer segmentation.

Cost-utility analysis of inotuzumab ozogamicin for relapsed or refractory B cell acute lymphoblastic leukemia from the perspective of Taiwan's health care system

OBJECTIVES

We conduct a cost-utility analysis of inotuzumab ozogamicin (INO) versus chemotherapy as the standard of care (SOC) for adults with relapsed or refractory B cell acute lymphoblastic leukemia.

METHODS

A Markov model incorporating transition probabilities between health states was applied to simulate disease progression. The model inputs, including overall survival, progression-free survival, and utility parameters, were obtained from the INO-VATE ALL trial and literatures. The Taiwan Cancer Registry Database and the Health and Welfare Database were utilized to identify the patient cohort and medical costs from the perspective of National Health Insurance Administration. The lifetime medical costs (in 2017 US dollars), quality-adjusted life years (QALYs) gained, and associated incremental cost-effectiveness ratio (ICER) were the main study outcomes.

RESULTS

The lifetime medical costs for INO and SOC were $176,795 and $69,496, and the QALYs gained were 2.25 and 0.84, respectively. The ICER for INO versus SOC was $76,044 per QALY gained, which is slightly more than three times Taiwan's gross domestic product per capita (i.e., $73,224). Favorable economic results for INO versus SOC were found with an increased time horizon for model simulation, less discounting for the future benefit, and higher stem cell transplantation (SCT) rate after INO treatment; and among patients aged less than 55 years, with no SCT history, or in the first salvage treatment.

CONCLUSIONS

INO versus SOC has higher costs but is more effective. The use of INO is favorable for patients in the early treatment course and when more future benefit associated with INO is considered.

Comparing the effectiveness of different EGFR-TKIs in patients with EGFR mutant non-small-cell lung cancer: A retrospective cohort study in Taiwan

The study was to compare the effectiveness of different epidermal growth factor receptor-tyrosine kinase inhibitors (EGFR-TKIs) in patients with advanced non-small-cell lung cancer (NSCLC) and received EGFR-TKIs as first-line therapy. This retrospective cohort study was conducted using data from real-world settings. Patients with stage IIIB and IV NSCLC and first received gefitinib, erlotinib, or afatinib between 2011 and 2015 were included. The date of the first claim for EGFR-TKIs was set as the index date. Study endpoints were all-cause death and treatment failure that was defined when patients added on or switched to chemotherapy or terminal care. A total of 5,940 patients, including 3,982 (67.0%) receiving gefitinib, 1,207 (20.3%) receiving erlotinib, and 751 (12.7%) receiving afatinib, were eligible for this study. The 1-year overall survival (OS) rates for gefitinib, erlotinib, and afatinib groups were 74% (95% confidence interval [CI]: 72-75%), 75% (95% CI: 73-77%), and 80% (95% CI: 77-83%), respectively. Compared to gefitinib, afatinib was associated with a lower risk of all-cause death (adjusted hazard ratio [aHR] = 0.82, 95% CI: 0.72-0.93) but not erlotinib (aHR = 0.95, 95% CI: 0.86-1.05). Similar results were also found regarding the effectiveness of treatment. All the three EGFR-TKIs showed no differences for both outcomes among patients with an Eastern Cooperative Oncology Group Performance Score of 2. The real-world data exhibited afatinib was more likely to be used for younger patients in a better condition than other EGFR inhibitors, and observed prolonged OS and treatment effectiveness compared to gefitinib after performing a multivariate Cox regression analysis.

In situ chemically crosslinked injectable hydrogels for the subcutaneous delivery of trastuzumab to treat breast cancer

Recently, novel approaches for the delivery of therapeutic antibodies have attracted much attention, especially sustained release formulations. However, sustained release formulations capable of carrying a high antibody load remain a challenge for practical use. In this study, a novel injectable hydrogel composed of maleimide-modified γ-polyglutamic acid (γ-PGA-MA) and thiol end-functionalized 4-arm poly(ethylene glycol) (4-arm PEG-SH) was developed for the subcutaneous delivery of trastuzumab. γ-PGA-MA and 4-arm PEG-SH formed a hydrogel through thiol-maleimide reactions, which had shear-thinning properties and reversible rheological behaviors. Moreover, a high content of trastuzumab (>100 mg/mL) could be loaded into this hydrogel, and trastuzumab demonstrated a sustained release over several weeks through electrostatic attraction. In addition, trastuzumab released from the hydrogel had adequate stability in terms of its structural integrity, binding bioactivity, and antiproliferative effect on BT-474 cells. Pharmacokinetic studies demonstrated that trastuzumab-loaded hydrogel (Her-hydrogel-10, composed of 1.5% γ-PGA-MA, 1.5% 4-arm PEG-SH, and 10 mg/mL trastuzumab) and trastuzumab/Zn-loaded hydrogel (Her/Zn-hydrogel-10, composed of 1.5% γ-PGA-MA, 1.5% 4-arm PEG-SH, 5 mM ZnCl₂, and 10 mg/mL trastuzumab) could lower the maximum plasma concentration (C_max) than the trastuzumab solution. Furthermore, Her/Zn-hydrogel-10 was better able to release trastuzumab in a controlled manner, which was ascribed to electrostatic attraction and formation of trastuzumab/Zn nanocomplexes. In a BT-474 xenograft tumor model, Her-hydrogel-10 had a similar tumor growth-inhibitory effect as that of the trastuzumab solution. By contrast, Her/Zn-hydrogel-10 exhibited a superior tumor growth-inhibitory capability due to the functionality of Zn. This study demonstrated that this hydrogel has potential as a carrier for the local and systemic delivery of proteins and antibodies. STATEMENT OF SIGNIFICANCE: Recently, novel sustained-release formulations of therapeutic antibodies have attracted much attention. However, these formulations should be able to carry a high antibody load owing to the required high dose, and these formulations remain a challenge for practical use. In this study, a novel injectable chemically cross-linked hydrogel was developed for the subcutaneous delivery of trastuzumab. This novel hydrogel possessed ideal characteristics of loading high content of trastuzumab (>100 mg/mL), sustained release of trastuzumab over several weeks, and maintaining adequate stability of trastuzumab. In vivo studies demonstrated that a trastuzumab-loaded hydrogel possessed the ability of controlled release of trastuzumab and maintained antitumor efficacy same as that of trastuzumab. These results implied that a γ-PGA-MA and 4-arm PEG-SH-based hydrogel has great potential in serving as a carrier for the local or systemic delivery of therapeutic proteins or antibodies.

Statin use and risk of COPD exacerbation requiring hospitalization

BACKGROUND

Despite recent studies that suggested statins' beneficial effects on chronic obstructive pulmonary disease (COPD) outcomes, the impact, if any, of statins on COPD exacerbations remains unclear. This study aimed to examine the association between statin use and risk of hospitalized COPD exacerbation, and to assess whether the association varied by statin initiation, dose, or duration of use.

METHODS

A retrospective nested case-control study among patients with COPD was conducted analyzing a nationwide health insurance claims database in Taiwan. Cases were subjects hospitalized for COPD exacerbations; each case was matched to 4 randomly selected controls on age, sex, cohort entry, and number of COPD-related outpatient visits by an incident-density sampling approach. Conditional logistic regressions were employed to quantify the COPD exacerbation risk associated with statin use.

RESULTS

The study cohort comprised 14,316 COPD patients, from which 1584 cases with COPD exacerbations and 5950 matched controls were identified. Any use of statins was associated with a 30% decreased risk of COPD exacerbation (95% confidence interval [CI], 0.56-0.88), and current use of statins was related to a greater reduced risk (adjusted odds ratio [OR] 0.60; 95% CI, 0.44-0.81). A dose-dependent reduced risk of COPD exacerbation by statins was observed (medium average daily dose: adjusted OR 0.60; 95% CI, 0.41-0.89; high daily dose: adjusted OR 0.33; 95% CI, 0.14-0.73). The reduced risk remained significant for either short or long duration of statin use.

CONCLUSIONS

Statin use was associated with a reduced risk of COPD exacerbation, with a further risk reduction for statins prescribed more recently or at high doses.

Analysis of Excess Direct Medical Costs of Vision Impairment in Taiwan

OBJECTIVES

To quantify the annual excess direct medical costs of vision impairment from the perspective of the Bureau of National Health Insurance in Taiwan and to examine whether the costs vary by severity and duration of vision impairment.

METHODS

A retrospective matched cohort analysis was conducted by using data from the Longitudinal Health Insurance Databases between January 1, 2000, and December 31, 2008. All patients newly diagnosed with vision impairment were categorized as having moderate vision loss, severe vision loss, or blindness. Each patient with vision impairment was matched to one randomly selected patient with normal vision by age (±1 year) and sex. At each level of vision impairment, generalized linear models were used to quantify the total annual excess costs and component costs incurred in the first and second years.

RESULTS

Vision impairment was associated with significantly higher crude excess medical costs. At each level of vision impairment, the total crude medical costs were attributable to different resource utilization and dominated by non-eye-related medical care. After adjusting for covariates, the first-year annual excess costs increased with escalating severity of vision impairment: New Taiwan (NT) $9894 for moderate vision loss, NT $22,760 for severe vision loss, and NT $52,687 for blindness. Similarly, the second-year adjusted costs were estimated as NT $3477, NT $19,532, and NT $28,272 for moderate vision loss, severe vision loss, and blindness, respectively.

CONCLUSIONS

Consistent with Western countries, vision impairment is associated with significantly increased health care costs in Taiwan. The excess costs seem to increase with severity of vision impairment and decrease in the second year.

Risk of stroke associated with inhaled ipratropium bromide in chronic obstructive pulmonary disease: a population-based nested case-control study

BACKGROUND

Cardiovascular safety concerns about inhaled ipratropium bromide have recently been raised. Nonetheless, the specific stroke risk associated with ipratropium use has not been evaluated thoroughly.

METHODS

This was a population-based nested case-control study analyzing data from the National Health Insurance Research Database in Taiwan. A cohort of 15,396 newly-diagnosed chronic obstructive pulmonary disease (COPD) patients was included between 2001 and 2007, in which 1477 cases of incident hospitalization for stroke were identified. Each case was individually matched to four randomly-selected controls based on age, sex, and cohort entry date. Conditional logistic regressions were used to estimate the odds ratio (OR) for risk of stroke-related hospitalization associated with ipratropium use.

RESULTS

Any use of ipratropium within the 6 months before the index date was associated with an increased risk of stroke compared with nonuse (adjusted OR, 2.02; 95% CI, 1.71 to 2.41). The observed risk remained significant regardless of accumulated doses. Additionally, use of ipratropium within 30 days before the index date resulted in the greatest risk (adjusted OR, 2.97 95% CI, 2.27 to 3.88). Furthermore, an increased risk of stroke was found for ipratropium regimens involving concomitant use of inhaled short-acting β(2)-agonists (SABAs; adjusted OR, 2.18; 95% CI, 1.81 to 2.62) or theophyllines (adjusted OR, 1.79; 95% CI, 1.42 to 2.26).

CONCLUSIONS

Use of ipratropium is associated with an increased risk of stroke in COPD patients. Clinicians should be alert to that risk when prescribing ipratropium, especially for those receiving ipratropium more recently or those with concomitant use of SABAs or theophyllines.

Aortic atheroma is related to outcome but not numbers of emboli during coronary bypass

BACKGROUND

The relation between aortic atheroma severity and stroke after coronary artery bypass grafting is established. The relation between atheroma severity and other outcome measures or numbers of emboli has not been determined.

METHODS

Using transesophageal echocardiography, we determined the severity of atheroma in the ascending, arch, and descending aortic segments in 84 patients undergoing operations. Seventy patients were monitored using transcranial Doppler ultrasonography.

RESULTS

The incidence of stroke was 33.3% among 9 patients with mobile plaque of the arch and 2.7% among 74 patients with nonmobile plaque (p = 0.011). Cardiac complications were not significantly related to atheroma severity in any aortic segment. Length of stay was significantly related to atheroma severity in the aortic arch (p = 0.025) and descending segment (p = 0.024). The presence of severe atheroma in both the arch and descending segments was associated with significantly longer hospital stays as compared with patients with severe atheroma in neither segment (p = 0.05). Numbers of emboli were greater in patients with severe atheroma at clamp placement, although the differences did not achieve statistical significance.

CONCLUSIONS

Aortic atheroma severity is related to stroke and to the duration of hospitalization after coronary artery bypass grafting. The lack of correlation between numbers of emboli and atheroma severity suggests that m any emboli may be nonatheromatous in nature.

Determination of size of aortic emboli and embolic load during coronary artery bypass grafting

BACKGROUND

Embolic signals have been detected within both the aortic lumen and the intracranial vasculature during coronary artery bypass grafting. Total numbers of these emboli have been reported. The present study examined the size of individual emboli and the total volume of embolization.

METHODS

Using transesophageal echocardiography, we continuously monitored the aortic lumen of 10 patients undergoing isolated coronary artery bypass grafting. We manually analyzed 720,000 individual echo frames over a 4-minute period after the release of aortic clamps to track and to calculate the volume of 657 individual particles. The embolic load for the entire procedure was calculated from mean volume based on analysis of 1,508 particles. We simultaneously monitored the middle cerebral artery using transcranial Doppler ultrasonography and compared numbers of emboli detected by the two techniques.

RESULTS

Particle diameter ranged from 0.3 to 2.9 mm (mean, 0.8 mm), and particle volume from 0.01 to 12.5 mm3 (mean, 0.8 mm3). Twenty-eight percent of particles measured 1 mm or more, 44% measured 0.6 to 1.0 mm, and only 27% measured 0.6 mm or less in diameter. Aortic embolic load for the procedure ranged from 0.6 cm3 to 11.2 cm3 (mean, 3.7 cm3). Estimated cerebral embolic load for the procedure ranged from 60 to 510 mm3 (mean, 276 mm3). The fraction of aortic emboli entering the cerebral circulation was very variable (3.9% to 18.1%). Seventy-six percent of the embolic volume after the release of clamps occurred over a 20-second period. Only 1 patient was encephalopathic perioperatively. This patient had the largest estimated cerebral embolic load (510 mm3) and the second largest aortic embolic load (8.4 cm3).

CONCLUSIONS

We determined the size of individual intraaortic embolic particles and the total volume of embolization during coronary artery bypass grafting, and found the proportion entering the cerebral circulation to be very variable. The constitution of these particles and the neurologic impairment resulting from such embolization remains to be determined.

Impact of embolization during coronary artery bypass grafting on outcome and length of stay

BACKGROUND

Transcranial Doppler ultrasonography detects emboli in most patients during coronary artery bypass grafting. However, the significance of these emboli has not yet been established.

METHODS

We monitored 82 patients during coronary artery bypass grafting with this technique and related the numbers of emboli to the outcomes and length of hospital stay.

RESULTS

We detected cerebral emboli in all patients. Patients with stroke (n = 4; 4.9%) had a mean of 449 emboli, as compared with 169 emboli in patients without stroke (n = 78) (p = 0.005). Patients with major cardiac complications (n = 7) had a mean of 392 emboli, as compared with 163 in patients without such complications (n = 75) (p = 0.003). The mean hospital stay of survivors was 8.6 days in patients with fewer than 100 emboli (n = 40), 13.5 days in patients with 101 to 300 emboli (n = 23), 16.3 days in those with 301 to 500 emboli (n = 16), and 55.8 days in patients with more than 500 emboli (n = 6) (p = 0.0007). This relation was unchanged when patients with complications were excluded. The correlation between embolization and outcome was independent of the extent of aortic atheroma or age.

CONCLUSIONS

Emboli detected during coronary artery bypass grafting are significantly related to major cardiac and neurologic complications and affect length of stay in all patients, even in the absence of such specific complications.