Search

Consistent segmentation of CT scans in COVID-19 patients across multiple time points is important to accurately evaluate disease progression and therapeutic response. In medical domains, previous interactive segmentation studies have been mainly conducted on data from a single time point. However, the valuable segmentation information from previous time points is often underutilized in assisting the segmentation of a patient’s follow-up scans. Moreover, fully automatic segmentation techniques frequently produce results that would need further refinement for clinical applicability. In this study, we propose a novel single-network model for interactive segmentation that fully leverages all available past information to refine the segmentation of follow-up scans. In the first segmentation round, our model takes concatenated slices of 3D volumes from two-time points (target and reference), employing the segmentation results from the reference time point as a guide for segmenting the target scan. Subsequent refinement rounds incorporate user feedback in the form of scribbles that rectify the segmentation, in addition to incorporating the previous segmentation results of the target scan. This iterative process ensures the preservation of segmentation information from prior refinement rounds. Experimental results obtained from our in-house multiclass longitudinal COVID-19 dataset demonstrate the effectiveness of the proposed method compared to its static counterpart, thus providing valuable assistance in localizing COVID-19 infections in patients’ follow-up scans.

Purpose: The internet is an important information source for hand surgery fellowship applicants. A previous analysis of hand fellowship websites in 2014 demonstrated they were often inaccessible and incomplete. Given the increased importance of virtual information, we performed an updated assessment of the accessibility and content of hand fellowship program websites. Methods: Websites of 92 accredited hand surgery fellowship programs were evaluated for the following: (1) accessibility; and (2) the presence of 13 fellow recruitment and 13 fellow education criteria, as defined in prior studies. We used Mann-Whitney U and Kruskal-Wallis tests to assess whether the geographic region, number of fellows, or affiliation with a top orthopedic hospital or medical school were associated with website content. Results: Functional website links that redirect to the appropriate fellowship program website are provided for 47 (51.1%) of 92 programs on the American Society for Surgery of the Hand fellowship directory. All missing websites were accessible via independent Google searches. Fellowship program websites contained an average of 13.9 ± 4.4 total criteria (range, 3–23). Of the 15 criteria examined in both 2014 and 2021, there were significant (P < .05) increases in the prevalences of 4: current fellow(s), salary, social media links, and operative experience. Conclusions: Despite a slight increase in accessibility since 2014, nearly half of hand surgery fellowship program websites remain inaccessible from the American Society for Surgery of the Hand directory. Program websites averaged approximately half of the criteria analyzed, with many websites failing to provide information deemed important by applicants. Clinical relevance: Our study provides an impetus for improving the accessibility and content of hand surgery fellowship websites. A website that incorporates criteria examined in this study can serve as an effective recruitment tool by providing consistent baseline information that may help applicants decide which programs align with personal values and future career goals.

Purpose: To systematically review the literature for studies investigating the biomechanical properties of constructs used to repair isolated subscapularis tears in time zero human cadaveric studies. Methods: A systematic review was performed using Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Three electronic databases were searched for studies that reported on the construct technique and biomechanical outcomes for the repair of isolated subscapularis tears in human cadaveric specimens. Ultimate load, gap formation, stiffness, and failure mode were documented. Methodological quality was assessed using the Quality Appraisal for Cadaveric Studies (QUACS) scale. Results: Six articles qualified (104 shoulders [72 single-row, 26 double-row, 6 transosseous]; mean QUACS score 10.5 ± 1) and were analyzed. Studies varied in the number and type of anchors and construct technique (1-2 anchors single-row; 3-4 anchors double-row; bioabsorbable or titanium anchors) and suture(s) used (no. 2 FiberWire or FiberTape), subscapularis tear type (25%, 33%, 50%, or 100% tear), and whether a knotless or knotted fixation was used. In studies that created full-thickness, upper subscapularis tears (Fox-Romeo II/III or Lafosse II), no significant differences were seen in ultimate load, gap formation, and stiffness for knotted versus knotless single-row repair (2 studies) and single-row versus double-row repair (1 study). Double-row repair of complete subscapularis tears demonstrated higher ultimate load, stiffness, and lower gap formation in 1 study. Ultimate load differed between the studies and constructs (single-row: range, 244 N to 678 N; double-row: range 332 N to 508 N, transosseous: 453 N). Suture cutout was the most common mode of failure (59%). Conclusion: Because of the limited number of studies and varying study designs in examining the biomechanical properties of repair constructs used for subscapularis tears, there is inconclusive evidence to determine which construct type is superior for repairing subscapularis tears. Clinical Relevance: Results from biomechanical studies of clinically relevant subscapularis repair constructs are important to guide decision-making for choosing the optimal construct for patients with subscapularis tears.

Motivated by our collaboration with Anheuser-Busch InBev (AB InBev), a consumer packaged goods (CPG) company, we consider the problem of forecasting sales under the coronavirus disease 2019 (COVID-19) pandemic. Our approach combines nonparametric regression, game theory, and pandemic modeling to develop a data-driven competitive online nonparametric regression method. Specifically, the method takes the future COVID-19 case estimates, which can be simulated via the susceptible-infectious-removed (SIR) epidemic model as an input, and outputs the level of calibration for the baseline sales forecast generated by AB InBev. In generating the calibration level, we focus on an online learning setting where our algorithm sequentially predicts the label (i.e., the level of calibration) of a random covariate (i.e., the current number of active cases) given past observations and the generative process (i.e., the SIR epidemic model) of future covariates. To provide robust performance guarantee, we derive our algorithm by minimizing regret, which is the difference between the squared ℓ2 -norm associated with labels generated by the algorithm and labels generated by an adversary and the squared ℓ2 -norm associated with labels generated by the best isotonic (nondecreasing) function in hindsight and the adversarial labels. We develop a computationally efficient algorithm that attains the minimax-optimal regret over all possible choices of the labels (possibly non-i.i.d. and even adversarial). We demonstrate the performances of our algorithm on both synthetic and AB InBev's data sets of three different markets (each corresponds to a country) from March 2020 to March 2021. The AB InBev's numerical experiments show that our method is capable of reducing the forecast error in terms of weighted mean absolute percentage error (WMAPE) and mean squared error (MSE) by more than 37% for the company. This paper was accepted by J. George Shanthikumar, data science. Funding: This work was partially supported by the Massachusetts Institute of Technology Data Science Lab and AB-InBev Corporation. Supplemental Material: The e-companion and data are available at https://doi.org/10.1287/mnsc.2023.4969. [ABSTRACT FROM AUTHOR]

Abstract Sexual abuse against women and girls in Malawi is pervasive, and survivors face significant barriers in their quest for justice. One particular barrier--the "corroboration rule"--stands out as a discriminatory [...]

Host-generated property images as a visual channel reveal substantial information about properties. Selecting proper images to display can lead to higher demand and increased rental revenue. In this paper, we define, estimate, and optimize the impacts of Airbnb photos on customers' renting decisions. We apply ResNet-50, a convolutional neural network model, to build two separate, supervised learning models to evaluate the image quality and room types posted by Airbnb hosts. Then, we characterize the overall impacts of photo layout by the room type featured in the photo, photo quality, and order of display on the listings' web pages. To address two estimation challenges in the Airbnb setting, namely, censored demand and changing consideration sets, we propose a novel pairwise comparison model that utilizes customers' booking sequence data to consistently estimate the impact of photo layout on customers' renting decisions. Our estimation results suggest that the cover image has a significantly larger impact than noncover photos and a high-quality bedroom cover image leads to the largest increase in demand. Furthermore, we build a nonlinear integer programming optimization problem and develop an algorithm to determine the optimal photo layout. Our counterfactual analysis suggests that a listing's unilateral adoption of optimal photo layout leads to 11.0% more bookings on average. Moreover, depending on the neighborhood and market size, when listings simultaneously switch to the optimal photo layout, they get booked for two to five additional days in a year on average, which boosts revenue by $500 to $1,100. This paper was accepted by Swaminathan, Jayashankar, operations management. Funding: This research was partially sponsored by the MIT Data Science Lab and also benefited from generous support provided by Zalando. Supplemental Material: The online companion and data are available at https://doi.org/10.1287/mnsc.2022.4616. [ABSTRACT FROM AUTHOR]

Abstract. The presence of strategic customers may force an already financially distressed firm into a death spiral: sensing the Arm's financial difficulty, customers may wait strategically for deep discounts in [...]

Advances in Information and Communication Technology (ICT) and the distributed nature of innovation in a culturally heterogeneous world raise an important question about how ICT and culture jointly influence innovation. We argue that ICT shapes the influence of culture dimensions on national innovation in nuanced ways, depending on the nature of the dimensions. In particular, cultural dimensions that pertain to norms, formality, and structure (e.g., power distance and uncertainty avoidance) are often considered detrimental to innovation, whereas culture dimensions that pertain to the pursuit of goals and achievement (e.g., individualism, masculinity, and long-term orientation) are often considered conducive for innovation. We hypothesize that ICT amplifies the positive effects of individualism, masculinity, and long-term orientation on national innovation—and, attenuates the negative effects of power distance and uncertainty avoidance on national innovation. Our analyses of data on 66 countries from 2007 to 2015 broadly support our hypotheses about ICT amplifying the positive effects of individualism and long-term orientation, as well as our hypotheses about ICT attenuating the negative effects of power distance and uncertainty avoidance on national innovation. The findings suggest that ICT amplifies the effects of culture dimensions related to the pursuit of goals and helps overcome barriers to innovation that arise from a culture's focus on norms, formality, and structure. In turn, these findings can help managers and policy makers to make judicious decisions regarding ICT investments to transcend or leverage the country's cultural characteristics for innovation. [ABSTRACT FROM AUTHOR]

Retailers face significant pressure to improve revenue, margins and market share by applying price optimisation models. These are mathematical models that calculate how demand varies at different price levels, then combine that data with information on costs and inventory levels to recommend prices that will improve revenue and profits. These models have been around for a while-so what is different now? We have identified three important changes: (1) Data: availability of internal and external real-time data such as traffic to a website, consumers making buy/no buy decisions and competitor pricing strategies; (2) Analytics: advances in machine learning and ease of access (R, Python) have enabled the development of systems that learn on the fly about consumer behaviour and preferences and generate effective estimates of demand-price relationships; and (3) Automation: increase in computing speed enables real-time optimisation of prices of hundreds of competing products sold by the same retailer. We take advantage of these new opportunities by showing how they were applied at Boston-based flash sales retailer Rue La La, online market maker Groupon, and the largest online retailer in Latin America, B2W Digital (B2W). While all these examples are of on-line businesses which have readily available data and can change prices dynamically, we have also implemented similar methods for brick-and-mortar retailed in applications such as promotional pricing, new product introduction, and assortment optimisation with similar business impacts. Thus, beyond applications to price optimisations, these new trends enable companies to revolutionise their business from procurement to supply chain all the way to revenue management. [ABSTRACT FROM AUTHOR]