From ChatGPT to Treatment: the Future of AI and Large Language Models in Surgical Oncology

This paper explores the transformative potential of Large Language Models (LLMs) within the context of surgical oncology and outlines the foundational mechanisms behind these models. LLMs, such as GPT-4, have rapidly evolved in terms of scale and capabilities, with profound implications for their applications in healthcare. These models, rooted in the Generative Pretrained Transformer architecture, exhibit advanced natural language understanding and generation skills. Within surgical oncology, LLMs, when integrated into a Generalist Medical AI (GMAI) framework, hold great promise in offering real-time support throughout the cancer journey. However, alongside these opportunities, this paper underscores the importance of ethical, privacy, and efficacy considerations, especially in light of issues like data drift and potential biases. Collaborative efforts among healthcare providers, AI developers, and regulatory bodies are pivotal in ensuring responsible and effective use of LLMs in surgical oncology, thereby contributing to enhanced patient care and safety. As LLMs continue to advance, they are poised to become indispensable tools in the delivery of high-quality, efficient care in this specialized medical field.

Transcarotid Approach to Placement of an Impella 5.0

Microaxial left ventricular assist devices (mLVADs) have traditionally been placed through a transfemoral or transaxillary arterial approach. Transfemoral access is restrictive, significantly limiting postoperative patient ambulation. Transaxillary placement is preferred but not feasible in a subset of patients due to small arterial diameter or tight angulation of the thoracic outlet. Transcarotid delivery has been utilized for other cardiovascular device deployment with good success; however, this approach has not been described for mLVAD support. We present a case series of transcarotid placement of mLVADs in cases where a transaxillary and transfemoral approach was not feasible. From May 2017 to April 2019, six patients in cardiogenic shock required mLVAD support achieved via a transcarotid approach. Technical success was achieved in all patients. One patient was directly weaned from mLVAD support and two patients died on mLVAD support. Escalation to venoarterial extracorporeal membrane oxygenation (VA-ECMO) was required for three patients, two of whom subsequently died. There were no bleeding or valvular complications related to device placement, and no obvious or known neurologic complications related to mLVAD support. Transcarotid placement of mLVADs expands the utility of these devices as an alternative to traditional support strategies or prohibitive arterial anatomy; however, further study is needed to determine its efficacy.

Seasonal trends in donor heart availability: an analysis of the UNOS database

Despite the widespread belief that donor organ availability varies around holidays and seasons, there is little empirical data supporting this long-held belief. Variations in donor heart availability may be of interest to patients and clinicians. The UNOS/OPTN registry was queried for all heart donations from October 1987 through March 2017. Daily heart donation rates were modeled nationally using Poisson regression including splines for year and day of the year. Seasonality was assessed using a likelihood ratio test for the spine terms for day of the year. The holiday effect was assessed using conditional logistic regression. Seasonal plots suggest a significant, although modest, increase in organ availability during the summer months, except for region 1. The regions with the highest amplitude were region 7 (peak: June 21, amplitude: 16.63%) and region 6 (peak: July 5, amplitude: 11.29%). There was no significant difference in the odds of heart donation when comparing holidays vs. non-holidays using national data (odds ratio [95% CI]: 1.01 [0.98, 1.03], P = 0.560) or any regional subsets. There was no observable correlation between donor heart availability and holidays. However, a significant seasonality effect was observed with higher donation rates occurring during warmer months.

Temporary mechanical circulatory support prevents the occurrence of a low-output state in high-risk coronary artery bypass grafting: A case series

BACKGROUND

Coronary artery bypass grafting (CABG) is a durable treatment for coronary artery disease. Left ventricular dysfunction (LVD) (a division of cardiothoracic surgery) (ejection fraction < 35%) significantly elevates perioperative risk for patients pursuing surgical revascularization. Periprocedural support with temporary mechanical circulatory support (tMCS) has shown benefit in this patient population.

METHODS

Four patients with ischemic cardiomyopathy and LVD underwent CABG at our institution between 2017 and 2018. Each patient received perioperative ventricular support using a microaxial tMCS device (Impella 5.0®). The occurrence of a postoperative low-output state (LOS) was assessed for as well as postoperative morbidity and mortality, device-specific complications, and tMCS support duration.

RESULTS

All patients survived to device explant without device-related complications. Two patients required reoperation for nondevice-related bleeding. All patients were without an LOS at 24 h postoperatively with cardiac indices of 2.9-3.6 L/min/m² , normalized serum lactate, and vasoactive-inotrope scores of 0-12.0. There was a notably high incidence of acute renal failure (50%), which was observed in patients with preoperative cardiogenic shock. One patient died 10 days after the device explant. Of the three patients that survived to discharge, two were alive at the most recent follow-up. Postoperative device support varied widely (0-500 h).

CONCLUSION

Perioperative tMCS may be a viable strategy for preventing postoperative LOS in high-risk CABG patients with a low complication rate and acceptable morbidity. The application of microaxial tMCS devices in CABG is an area that warrants further investigation to delineate its impact on perioperative outcomes and potentially expand the indications for such devices.

CCR10 expression is required for the adjuvant activity of the mucosal chemokine CCL28 when delivered in the context of an HIV-1 Env DNA vaccine

An effective prophylactic vaccine targeting HIV must induce a robust humoral response and must direct the bulk of this response to the mucosa-the primary site of HIV transmission. The chemokine, CCL28, is secreted by epithelial cells at mucosal surfaces and recruits' cells expressing its receptor CCR10. CCR10 is predominantly expressed by IgA + ASCs. We hypothesized that co-immunization with plasmid DNA encoding consensus envelope antigens with plasmid-encoded CCL28 would enhance anti-HIV IgA responses at mucosal surfaces. Indeed, animals receiving pCCL28 and pEnvA/C had significantly increased HIV-specific IgA in fecal extract. Surprisingly, CCL28 co-immunization induced a significant increase in anti-HIV IgG in the serum in mice compared to those receiving pEnvA/C alone. These robust antibody responses were not associated with changes in the frequency of germinal center B cells but depended upon the expression of CCR10, as these responses we abolished in CCR10-deficient animals. Finally, immunization with CCL28 led to increased frequencies in HIV-specific CCR10 + and CCR10 + IgA + B cells in the small intestine and Peyer's patches of vaccinated animals as compared to those receiving pEnvA/C alone. These data indicate that CCL28 administration can enhance antigen-specific humoral responses systemically and at mucosal surfaces.

Induced Regenerative Elastic Matrix Repair in LOXL1 Knockout Mouse Cell Cultures: Towards Potential therapy for Pelvic Organ Prolapse

Impaired elastic matrix remodeling occurs in reproductive tissues after vaginal delivery. This has been linked to development of pelvic organ prolapse (POP) for which there currently is no pharmacologic therapy. Hyaluronan oligomers and transforming growth factor beta 1 (termed elastogenic factors, EFs) have been shown to significantly enhance tropoelastin synthesis, elastic fiber assembly, and crosslinking by adult vascular smooth muscle cells (SMCs). The goal of this study was to ascertain if these factors similarly improve the quantity and quality of elastic matrix deposition by vaginal SMCs (VSMCs) isolated from lysyl oxidase like-1 knock out (LOXL1 KO) mouse model of POP. Cells isolated from whole vagina of a LOXL1 KO mouse (multiparous, stage 3 prolapse) were cultured and identified as SMCs by their expression of various SMC markers. Passage 2 vaginal SMCs (VSMCs; 3×10⁴/10 cm²) were cultured for 21 days with EFs. Cell layers and spent medium aliquots were assessed for elastin content and quality. EF-treated VSMCs proliferated at a similar rate to untreated controls but synthesized more total elastin primarily in the form of soluble matrix elastin. Elastin mRNA was also increased compared to controls. The elastic matrix was significantly denser in EF-treated cultures, which was composed of more mature, non-interrupted elastic fibers that were absent in controls. The results are promising towards development of a therapy to enhance regenerative elastic matrix repair in post-partum female pelvic floor tissues.

Characterization of glycidyl methacrylate - crosslinked hyaluronan hydrogel scaffolds incorporating elastogenic hyaluronan oligomers

Prior studies on two-dimensional cell cultures suggest that hyaluronic acid (HA) stimulates cell-mediated regeneration of extracellular matrix structures, specifically those containing elastin, though such biologic effects are dependent on HA fragment size. Towards being able to regenerate three-dimensional (3-D) elastic tissue constructs, the present paper studies photo-crosslinked hydrogels containing glycidyl methacrylate (GM)-derivatized bio-inert high molecular weight (HMW) HA (1 × 10(6)Da) and a bioactive HA oligomer mixture (HA-o: MW ∼0.75 kDa). The mechanical (rheology, degradation) and physical (apparent crosslinking density, swelling ratio) properties of the gels varied as a function of incorporated HA oligomer content; however, overall, the mechanics of these hydrogels were too weak for vascular applications as stand-alone materials. Upon in vivo subcutaneous implantation, only a few inflammatory cells were evident around GM-HA gels, however their number increased as HA-o content within the gels increased, and the collagen I distribution was uniform. Smooth muscle cells (SMC) were encapsulated into GM hydrogels, and calcein acetoxymethyl detection revealed that the cells were able to endure twofold the level of UV exposure used to crosslink the gels. After 21 days of culture, SMC elastin production, measured by immunofluorescence quantification, showed HA-o to increase cellular deposition of elastic matrix twofold relative to HA-o-free GM-HA gels. These results demonstrate that cell response to HA/HA-o is not altered by their methacrylation and photo-crosslinking into a hydrogel, and that HA-o incorporation into cell-encapsulating hydrogel scaffolds can be useful for enhancing their production of elastic matrix structures in a 3-D space, important for regenerating elastic tissues.

Impact of delivery mode of hyaluronan oligomers on elastogenic responses of adult vascular smooth muscle cells

Our prior studies demonstrated that exogenous supplements of pure hyaluronan (HA) tetramers (HA4) dramatically upregulate elastin matrix synthesis by adult vascular smooth muscle cells (SMCs). Some studies suggest that exogenous HA likely only transiently contacts and signals cells, and may elicit different cell responses when presented on a substrate (e.g., scaffold surface). To clarify such differences, we used a carbodiimide-based chemistry to tether HA4 onto glass, and compared elastin matrix synthesis by SMCs cultured on these substrates, with those cultured with equivalent amounts of exogenous HA4. Tethered HA4-layers were first characterized for homogeneity, topography, and hydrolytic stability using SEM, XPS, AFM, and FACE. In general, mode of HA4 presentation did not influence its impact on SMC proliferation, or cell synthesis of tropoelastin and matrix elastin, relative to non-HA controls; however, surface-tethered HA4 stimulated SMCs to generate significantly greater amounts of elastin-stabilizing desmosine crosslinks, which partially accounts for the greater resistance to enzymatic breakdown of elastin derived from these cultures. Elastin derived from both sets of cultures contained peptide masses that correspond to the predominant peptides present in rat aortic elastin. SEM and TEM showed that HA4-stimulated fibrillin-mediated elastin matrix deposition, and organization into fibrils. Surface-immobilized HA4 was particularly conducive to organization of elastin into aggregating fibrils, and their networking to form closely woven sheets of elastin fibers, as seen in cardiovascular tissues. The results suggest that incorporation of elastogenic HA4 mers onto cell culture substrates or scaffolds is a better approach than exogenous supplementation for in vitro or in vivo regeneration of architecturally and compositionally faithful-, and more stable mimics of native vascular elastin matrices.

Evaluation of the matrix-synthesis potential of crosslinked hyaluronan gels for tissue engineering of aortic heart valves

Our goal is to fabricate continuous sheets of elastin atop non-biodegradable hydrogels (hylans) containing crosslinked hyaluronan, a glycosaminoglycan. Such elastin-hyaluronan composites may be useful to tissue engineer replacements for the glycosaminoglycan- and elastin-rich layers of the native aortic valve cusp. Neonatal rat aortic smooth muscle cells were cultured atop hylan gels with micro-textured surfaces, and on plastic, and the components of the extracellular matrix (collagen, elastin) were periodically analyzed. The hylan substrates induced the cells to proliferate more rapidly and over longer time periods (approximately 4 weeks) relative to those cultured on plastic (2-3 weeks). Consequently, at all assay times, the amounts of elastin was derived from the hylan-based cell cultures was 25% or more than that derived from cells cultured on plastic. However, when elastin content was normalized to the cell DNA content, no significant differences were found in the two substrates beyond the first two weeks of culture. Conversely, at culture times greater than 2 weeks, cells cultured atop hylan gels produced amounts of collagen/nanogram of DNA that were approximately 56% less than that synthesized by cells cultured on plastic. Cells grown on hylan deposited an unusual matrix layer, rich in elastin, at the hylan-cell interface. This elastin was found to be organized into fenestrated sheets and loose elastin fibers, structures that were also isolated from the elastin matrix of the ventricularis layer of porcine aortic valve cusps. We have thus demonstrated that hylan gels are useful as substrates to induce elastin synthesis in culture to obtain structures that resemble the elastin matrix of the native aortic valve.

Effects of Nitric Oxide (NO) and Soluble Nucleoside Triphosphate Diphosphohydrolase (NTPDase) on Inhibition of Platelet Deposition In Vitro

Vascular thrombosis is regulated via the release of several constituents from the vascular endothelium, including nucleoside triphosphate diphosphohydrolases (NTPDases or ectonucleotidases), nitric oxide (NO), and eicosanoids. Currently, it is unknown how these constituents interact in the inhibition of platelet aggregation and adhesion. To investigate the combined effects of NO and NTPDase on platelet deposition sequestration, an in vitro study was performed to compare inhibition of platelet deposition to a biomaterial by NO in the absence or presence of soluble NTPDase. Results of the platelet inhibition studies with NO and NTPDase conclusively show that the inhibitory effects of NTPDase and NO are additive. The platelet inhibitory potency in the presence of NO was enhanced by NTPDase in a dose-dependent manner, for a given NO exposure. This augmentation is independent of aspirin; the ability of NTPDase or NO alone to inhibit platelet deposition is also independent of aspirin. Clearly, NO and NTPDase independently contribute to platelet inhibition via different mechanisms. The inaction of NO on the activity of NTPDase confirmed that NO or reaction products in the presence of O(2) do not interact with NTPDase directly.

Nitric oxide inhibition of platelet deposition on biomaterials

Platelet adhesion and aggregation restrict the clinical applicability of blood-contacting biomaterials. Nitric oxide (NO) is a simple biological molecule that may be incorporated into biomaterials to inhibit platelet deposition. The toxicity of NO at superphysiological levels necessitates the determination of aqueous NO concentrations and fluxes that effectively inhibit platelet deposition. In this study, a novel NO delivery device has been developed to study NO inhibition of platelet deposition in a dynamic in vitro environment. Gaseous NO was delivered via a semipermeable membrane to a radiolabeled platelet suspension perfusing a thin flow slit. The membrane was coated with a platelet-agonistic protein. Spatial NO flux and concentration profiles in the flow slit are predictable using a mathematical model. Platelet inhibition was essentially complete at 0.1 ppm gaseous NO exposure, corresponding to a surface concentration of 0.09 nM and surface fluxes between 0.3 and 0.6 femtomoles cm-2s-1. These threshold values of NO exposure for significant platelet inhibition were unchanged irrespective of the platelet agonist, perfusion times, or shear rates. At lower NO exposures (0.02 ppm), platelet inhibition was only partial with the degree of inhibition dependent on the nature of the agonistic protein. This study yields information useful towards the design and development of biomaterials incorporating NO for the reduction of platelet-biomaterial interactions.

Influence of agonist, shear rate, and perfusion time on nitric oxide inhibition of platelet deposition

Nitric oxide (NO) is a physiological species involved in inhibition of platelet adhesion and aggregation. A novel NO delivery device was utilized to quantitatively assess the effects of gaseous NO on platelet deposition to agonist-coated biomaterials in the presence of a platelet suspension. Platelet deposition was evaluated as a function of agonist (collagen, fibrinogen, or IgG), shear rate (250, 500, and 750 s(-1)), and perfusion time (5, 7.5, and 15 min). The minimal aqueous surface NO concentrations and fluxes necessary for significant inhibition of platelet deposition were quantified. Platelet deposition was completely inhibited at a gaseous NO exposure of 0.1 ppm, irrespective of the platelet agonist, shear rate, and perfusion time. The corresponding aqueous surface NO concentration was 0.09 nM at 250 s(-1) as predicted by a validated model. Surface fluxes ranged between 0.3 and 0.6 femtomoles cm(-2) s(-1). The results of this study are useful for establishing generalized guidelines (i.e., NO flux requirements in the presence of agonists, shear rate, and perfusion time) for the design and development of suitable biomaterials incorporating NO to reduce platelet deposition. Further studies incorporating blood, rather than platelet suspensions, are required to provide a more complete assessment of the required NO flux necessary to inhibit platelet deposition.

Design of a novel apparatus to study nitric oxide (NO) inhibition of platelet adhesion

Nitric oxide (NO) is a simple biological molecule which inhibits adhesion and aggregation of platelets. A novel NO delivery device has been developed to quantitatively study the effects of NO concentration and flux on the adhesion of platelets to a surface. The slit-flow device is lined with a protein-coated membrane through which NO gas permeates into a perfusing platelet suspension. A model predicting spatial NO concentrations and fluxes within the flow slit was validated. At a wall shear rate of 250 s(-1), platelet adhesion was inhibited 87% relative to controls for exposures as low as 0.1 ppm NO. Corresponding model predictions of the aqueous NO concentration and fluxes at the surface were 0.15 nM, and between 0.5 and 1.1 nanomoles cm(-2) s(-1), respectively. Endothelial cells, which release NO to inhibit platelet adhesion in vivo, generate NO at an estimated flux similar to the above values. At a NO exposure of 0.02 ppm, platelet inhibition was only 10%. The delivery device is useful for other studies in which a knowledge of the spatial NO fluxes or concentrations is desired. Knowledge of these fluxes or concentrations is beneficial towards the design of biomaterials incorporating NO to inhibit platelet adhesion.

Measurement and modeling of nitric oxide release rates for nitric oxide donors

An accurate model of the nitric oxide (NO)-release rate is essential for predicting the temporal NO-release rate and resulting NO concentrations for NO donors. Knowledge of the NO-release rate and/or the NO concentration is beneficial for assessing the physiological or pathological effects of NO on cell systems. This study describes a method to measure the temporal NO-release rate from NO donor compounds utilizing a modified ultrafiltration cell. For this study, the NO-release rates of spermine NONOate and diethylamine NONOate were measured and kinetically modeled at pH 7.4 and 37 degrees C. An advantage of this method is that complete dissolution of the NONOate was not necessary for modeling the NO-release rate. One model parameter, which is the number of moles of NO released per mole of decomposed NONOate, is 1.7 +/- 0.1 and 1.5 +/- 0.2 for spermine and diethylamine NONOate, respectively. The other model parameter, which is the NONOate first-order decomposition rate constant, is 0.019 +/- 0.002 min-1 and 0.47 +/- 0.10 min-1 for spermine and diethylamine NONOate, respectively, as determined from NO concentration profiles. The decomposition rate constant measured by spectrophotometry was consistent with the above value for spermine NONOate but was approximately half the above value for diethylamine NONOate. Preliminary experiments using spectrophotometry showed that for both NONOates the decomposition activation energy was approximately 100 kJ/mol. The NO-release rate model for spermine NONOate was applied to a model for predicting the NO concentration in oxygenated solution. The NO concentration was measured in phosphate buffer, culture medium, and Tyrode's solution. Excellent agreement was observed between experimental and predicted NO concentrations.