Deep learning for dermatologists: Part II. Current applications.

Because of a convergence of the availability of large data sets, graphics-specific computer hardware, and important theoretical advancements, artificial intelligence has recently contributed to dramatic progress in medicine. One type of artificial intelligence known as deep learning has been particularly impactful for medical image analysis. Deep learning applications have shown promising results in dermatology and other specialties, including radiology, cardiology, and ophthalmology. The modern clinician will benefit from an understanding of the basic features of deep learning to effectively use new applications and to better gauge their utility and limitations. In this second article of a 2-part series, we review the existing and emerging clinical applications of deep learning in dermatology and discuss future opportunities and limitations. Part 1 of this series offered an introduction to the basic concepts of deep learning to facilitate effective communication between clinicians and technical experts.

Deep learning for dermatologists: Part I. Fundamental concepts.

Artificial intelligence is generating substantial interest in the field of medicine. One form of artificial intelligence, deep learning, has led to rapid advances in automated image analysis. In 2017, an algorithm demonstrated the ability to diagnose certain skin cancers from clinical photographs with the accuracy of an expert dermatologist. Subsequently, deep learning has been applied to a range of dermatology applications. Although experts will never be replaced by artificial intelligence, it will certainly affect the specialty of dermatology. In this first article of a 2-part series, the basic concepts of deep learning will be reviewed with the goal of laying the groundwork for effective communication between clinicians and technical colleagues. In part 2 of the series, the clinical applications of deep learning in dermatology will be reviewed and limitations and opportunities will be considered.

How to Sequence Therapies in Mycosis Fungoides.

OPINION STATEMENT: Choice of therapy in mycosis fungoides is based on both patient- and lymphoma-specific factors, such as disease characteristics, comorbidities, symptoms and effect on quality of life, potential associated toxicities of therapy, response and tolerance to prior lines of therapy, and convenience and practicality. Generally, we sequence therapies from least toxic, targeted, nonimmunosuppressive to more toxic, immunosuppressive and from single agent to multiple agents, as necessary. If more toxic, immunosuppressive agents are required to alleviate disease burden or symptoms, we generally use them just long enough to control the disease, then transition to a maintenance regimen with less toxic, less immunosuppressive agents.

A microbial expression system for high-level production of scFv HIV-neutralizing antibody fragments in Escherichia coli.

Monoclonal antibodies (mABs) are of great biopharmaceutical importance for the diagnosis and treatment of diseases. However, their production in mammalian expression hosts usually requires extensive production times and is expensive. Escherichia coli has become a new platform for production of functional small antibody fragment variants. In this study, we have used a rhamnose-inducible expression system that allows precise control of protein expression levels. The system was first evaluated for the cytoplasmic production of super folder green fluorescence protein (sfGFP) in various production platforms and then for the periplasmic production of the anti-HIV single-chain variable antibody fragment (scFv) of PGT135. Anti-HIV broadly neutralizing antibodies, like PGT135, have potential for clinical use to prevent HIV transmission, to promote immune responses and to eradicate infected cells. Different concentrations of L-rhamnose resulted in the controlled production of both sfGFP and scFv PGT135 antibody. In addition, by optimizing the culture conditions, the amount of scFv PGT135 antibody that was expressed soluble or as inclusions bodies could be modulated. The proteins were produced in batch bioreactors, with yields of 4.9 g/L for sfGFP and 0.8 g/L for scFv. The functionality of the purified antibodies was demonstrated by their ability to neutralize a panel of different HIV variants in vitro. We expect that this expression system will prove very useful for the development of a more cost-effective production process for proteins and antibody fragments in microbial cells.

Protein costs do not explain evolution of metabolic strategies and regulation of ribosomal content: does protein investment explain an anaerobic bacterial Crabtree effect?

Protein investment costs are considered a major driver for the choice of alternative metabolic strategies. We tested this premise in Lactococcus lactis, a bacterium that exhibits a distinct, anaerobic version of the bacterial Crabtree/Warburg effect; with increasing growth rates it shifts from a high yield metabolic mode [mixed-acid fermentation; 3 adenosine triphosphate (ATP) per glucose] to a low yield metabolic mode (homolactic fermentation; 2 ATP per glucose). We studied growth rate-dependent relative transcription and protein ratios, enzyme activities, and fluxes of L. lactis in glucose-limited chemostats, providing a high-quality and comprehensive data set. A three- to fourfold higher growth rate rerouted metabolism from acetate to lactate as the main fermentation product. However, we observed hardly any changes in transcription, protein levels and enzyme activities. Even levels of ribosomal proteins, constituting a major investment in cellular machinery, changed only slightly. Thus, contrary to the original hypothesis, central metabolism in this organism appears to be hardly regulated at the level of gene expression, but rather at the metabolic level. We conclude that L. lactis is either poorly adapted to growth at low and constant glucose concentrations, or that protein costs play a less important role in fitness than hitherto assumed.

Lactococcus lactis†YfiA is necessary and sufficient for ribosome dimerization.

Dimerization and inactivation of ribosomes in Escherichia coli is a two-step process that involves the binding of ribosome modulation factor (RMF) and hibernation promotion factor (HPF). Lactococcus lactis MG1363 expresses a protein, YfiA(L) (l) , which associates with ribosomes in the stationary phase of growth and is responsible for dimerization of ribosomes. We show that full-length YfiA(L) (l) is necessary and sufficient for ribosome dimerization in L. lactis but also functions heterologously in vitro with E. coli ribosomes. Deletion of the yfiA gene has no effect on the growth rate but diminishes the survival of L. lactis under energy-starving conditions. The N-terminal domain of YfiA(L) (l) is homologous to HPF from E. coli, whereas the C-terminal domain has no counterpart in E. coli. By assembling ribosome dimers in vitro, we could dissect the roles of the N- and C-terminal domains of YfiA(L) (l) . It is concluded that the dimerization and inactivation of ribosomes in L. lactis and E. coli differ in several cellular and molecular aspects. In addition, two-dimensional maps of dimeric ribosomes from L. lactis obtained by single particle electron microscopy show a marked structural difference in monomer association in comparison to the ribosome dimers in E. coli.

Systematic identification of tRNAome and its dynamics in Lactococcus lactis.

Transfer RNAs (tRNAs) through their abundance and modification pattern significantly influence protein translation. Here, we present a systematic analysis of the tRNAome of Lactococcus lactis. Using the next-generation sequencing approach, we identified 40 tRNAs which carry 16 different post-transcriptional modifications as revealed by mass spectrometry analysis. While small modifications are located in the tRNA body, hypermodified nucleotides are mainly present in the anticodon loop, which through wobbling expand the decoding potential of the tRNAs. Using tRNA-based microarrays, we also determined the dynamics in tRNA abundance upon changes in the growth rate and heterologous protein overexpression stress. With a fourfold increase in the growth rate, the relative abundance of tRNAs cognate to low abundance codons decrease, while the tRNAs cognate to major codons remain mostly unchanged. Significant changes in the tRNA abundances are observed upon protein overexpression stress, which does not correlate with the codon usage of the overexpressed gene but rather reflects the altered expression of housekeeping genes.

Structural divergence of paralogous S components from ECF-type ABC transporters.

Energy coupling factor (ECF) proteins are ATP-binding cassette transporters involved in the import of micronutrients in prokaryotes. They consist of two nucleotide-binding subunits and the integral membrane subunit EcfT, which together form the ECF module and a second integral membrane subunit that captures the substrate (the S component). Different S components, unrelated in sequence and specific for different ligands, can interact with the same ECF module. Here, we present a high-resolution crystal structure at 2.1 Å of the biotin-specific S component BioY from Lactococcus lactis. BioY shares only 16% sequence identity with the thiamin-specific S component ThiT from the same organism, of which we recently solved a crystal structure. Consistent with the lack of sequence similarity, BioY and ThiT display large structural differences (rmsd = 5.1 Å), but the divergence is not equally distributed over the molecules: The S components contain a structurally conserved N-terminal domain that is involved in the interaction with the ECF module and a highly divergent C-terminal domain that binds the substrate. The domain structure explains how the S components with large overall structural differences can interact with the same ECF module while at the same time specifically bind very different substrates with subnanomolar affinity. Solitary BioY (in the absence of the ECF module) is monomeric in detergent solution and binds D-biotin with a high affinity but does not transport the substrate across the membrane.

Regulation of acetate kinase isozymes and its importance for mixed-acid fermentation in Lactococcus lactis.

Acetate kinase (ACK) converts acetyl phosphate to acetate along with the generation of ATP in the pathway for mixed-acid fermentation in Lactococcus lactis. The reverse reaction yields acetyl phosphate for assimilation purposes. Remarkably, L. lactis has two ACK isozymes, and the corresponding genes are present in an operon. We purified both enzymes (AckA1 and AckA2) from L. lactis MG1363 and determined their oligomeric state, specific activities, and allosteric regulation. Both proteins form homodimeric complexes, as shown by size exclusion chromatography and static light-scattering measurements. The turnover number of AckA1 is about an order of magnitude higher than that of AckA2 for the reaction in either direction. The Km values for acetyl phosphate, ATP, and ADP are similar for both enzymes. However, AckA2 has a higher affinity for acetate than does AckA1, suggesting an important role under acetate-limiting conditions despite the lower activity. Fructose-1,6-bisphosphate, glyceraldehyde-3-phosphate, and phospho-enol-pyruvate inhibit the activities of AckA1 and AckA2 to different extents. The allosteric regulation of AckA1 and AckA2 and the pool sizes of the glycolytic intermediates are consistent with a switch from homolactic to mixed-acid fermentation upon slowing of the growth rate.

Effect of iTRAQ labeling on the relative abundance of peptide fragment ions produced by MALDI-MS/MS.

The identification of proteins in proteomics experiments is usually based on mass information derived from tandem mass spectrometry data. To improve the performance of the identification algorithms, additional information available in the fragment peak intensity patterns has been shown to be useful. In this study, we consider the effect of iTRAQ labeling on the fragment peak intensity patterns of singly charged peptides from MALDI tandem MS data. The presence of an iTRAQ-modified basic group on the N-terminus leads to a more pronounced set of b-ion peaks and distinct changes in the abundance of specific peptide types. We performed a simple intensity prediction by using a decision-tree machine learning approach and were able to show that the relative ion abundance in a spectrum can be correctly predicted and distinguished from closely related sequences. This information will be useful for the development of improved method-specific intensity-based protein identification algorithms.

A time series analysis of immune checkpoint inhibitor use in the United States Medicare population: 2014-2019.

BACKGROUND: The adoption of immune checkpoint inhibitors (ICIs) has dramatically transformed the treatment of numerous cancers. Medicare is the largest payer in the US and pays for physician-administered drugs through its medical Part B benefit. The aim of this study was to describe trends in ICI utilization and corresponding government expenditures within the US Medicare population. METHODS: We analyzed Medicare data to describe trends in total number of claims, total annual expenditures, expenditures per patient, and expenditures per claim for ICIs from January 2014 to December 2019. RESULTS: From 2014 to 2019, utilization rates for each of the seven market approved ICIs in the US increased. Over this time period, total Medicare expenditure on ICIs increased 1916% from $285,506,498 to $5,755,319,571. Concurrently, overall Medicare Part B drug expenditure increased 57% from $23,679,547,748 to $37,271,080,631. Expenditures on ICIs accounted for 40% of the increase in total Medicare Part B drug spending over this time period. CONCLUSIONS: The rapid increase in utilization of ICIs has accounted for a disproportionate share of government drug spending growth in the United States. Policymakers can potentially curb spending growth by linking payments to patient outcomes.

Trends in Medicare utilization and reimbursement for wound debridement procedures 2012-2017.

Wound care accounts for more than $6 billion of health expenditure in the United States annually. In addition, Medicare reimbursement per relative value unit has remained flat with respect to inflation over the past 20 years. Yet, physicians face increasing operating costs. The economic effects of declining inflation-adjusted Medicare physician reimbursement on the use of wound debridement procedures have yet to be elucidated. This study describes trends in Medicare reimbursement, utilization, and total Medicare expenditure for wound debridement procedures. Using Medicare claims data, we aggregated the volume of services and average Medicare reimbursement for wound debridement procedures over the calendar years 2012 to 2017. For each billing code, we estimated inflation-adjusted total Medicare expenditure by multiplying the average inflation-adjusted Medicare reimbursement by the aggregate volume of services. The average inflation-adjusted payment for wound debridement procedures decreased by 6% from 2012 to 2017. Over this time period, the utilization of wound care procedures increased 32%. Total inflation-adjusted annual Medicare expenditure on wound care increased 24% from 2012 to 2017. Therefore, despite declining physician reimbursement rates, total Medicare expenditure on wound debridement increased from 2012 to 2017. This increase in total Medicare expenditure was driven by higher utilization of wound debridement procedures.

Quantifying the Financial Value of Clinical Specialty Choice and Its Association With Competitiveness of Admissions.

Background The factors influencing medical student clinical specialty choice have important implications for the future composition of the US physician workforce. The objective of this study was to determine the career net present values (NPVs) of US medical students' clinical specialty choices and identify any relationships between a specialty's NPV and competitiveness of admissions as measured by the US Medical Licensing Examination (USMLE) Step 1 scores. Methodology NPVs were calculated using the results of the 2019 Doximity Physician Compensation report, a survey of 90,000 physicians. Mean USMLE Step 1 scores for matched US allopathic seniors in the 2018 National Resident Matching Program were used as a measure of clinical specialties' competitiveness of admissions. We calculated a composite measure of NPV and annual work-hours by dividing each specialty's NPV by the reported average number of hours worked per year. Results In our analysis, orthopedic surgery had the highest NPV ($10,308,868), whereas family medicine had the lowest NPV ($5,274,546). Dermatology and plastic surgery had the highest mean USMLE Step 1 scores (249 for both), whereas family medicine had the lowest (220). Clinical specialties' NPVs were positively associated with mean USMLE Step 1 scores (Pearson's r = 0.82; p < 0.001). Conclusions In this study, we describe associations suggesting that medical students respond to financial incentives in choosing clinical specialties and that these decisions are mediated by USMLE Step 1 scores. This underscores the importance of titrating and aligning incentives to improve the allocation of medical students into clinical specialties.

The policy dimensions, regulatory landscape, and market characteristics of teledermatology in the United States.

The COVID-19 pandemic has spurred healthcare systems across the world to rapidly redesign their models of care delivery. As such, this pandemic has accelerated the adoption of teledermatology in the United States. However, it remains unknown whether this momentum will be maintained after the pandemic. The future of teledermatology in the United States will be significantly influenced by a complex set of policy, legal, and regulatory frameworks. An understanding of these frameworks will help dermatologists more effectively adopt and implement teledermatology platforms. In this article, we review the current state of teledermatology in the United States, including policy dimensions, the regulatory landscape, market characteristics, and future directions.