Predicting cardiovascular risk factors from facial & full body photography using deep learning

Introduction

The early and easy detection of pathological cardiovascular phenotypes can lead to an early medical intervention and thus slow or limit the development of cardiovascular diseases. As full body photographs are easily obtainable without the need of medical expertise, this modality holds the potential to be viable for screening of populations.

Purpose

Utilizing data from a population-based study, we examined the possibility to detect cardiovascular risk factors from total body photographs using deep learning.

Methods

A population-based cohort study was utilized. The first data release provides facial and full body photographs in dermatologic standard poses of 6500 participants (median age 62.0 years, 49.6% men) and corresponding cardiovascular risk factors. Here, we focus on the most prevalent ones: smoking status (prevalence: 19.0%), hypertension (35.3%) and diabetes (8.2%). Here we employ a 2D-Convolutional Resnet-18 Neural Network for predicting the risk factors. It receives the full body image, the facial image and age and sex as input. We compare this to a logistic regression model only including sex and age. Logistic Regression and Neural Network are employed in a 5-fold validation scheme and t-tested for significance.

Results

Our model provided a good detection of arterial hypertension (AUC 0.711, CI 0.684–0.739), while a logistic regression on age and sex yielded a significantly worse prediction (AUC 0.681, CI 0.679– 0.683, p<0.05). Additionally, it made a good detection of a positive smoking status (AUC 0.733, CI 0.711–0.754), being significantly better than a respective logistic regression on age and sex (AUC 0.598, CI 0.597–0.6, p<0.001). Lastly, it classified diabetes well (AUC 0.744, CI 0.724–0.764, p<0.001) in comparison to the logistic regression (AUC 0.681, CI 0.679–0.683, p<0.001).

Conclusion

The presence of cardiovascular risk factors can be detected from a total body photograph. As total body photographs can be easily obtained with a majority of digital cameras, including smart phones, this model represents a potentially widely applicable diagnostic tool to easily screen large parts of the population for relevant cardiovascular risk factors, making an early medical intervention possible.

Funding Acknowledgement

Type of funding sources: None.

AB0622 The Joint Vasculitis Registry in German-speaking countries (GeVas) – subgroup analysis of 113 GPA-patients

Background

Granulomatosis with polyangiitis (GPA) is the second most frequent vasculitis in Germany with an annual incidence of 34 per million and a prevalence of 210 per million [1]. GPA is characterized by its chronic course, frequent relapses, significant overall morbidity and mortality, and substantial socio-economic impact. Multiorgan involvement affecting the respiratory tract, kidney, and other organs is common. Limited variants also occur [2]. So far, prospective long-term observational data on the disease course of GPA are missing in Germany. Therefore, the Joint Vasculitis Registry in German-speaking countries (GeVas) has been established to follow the course of patients recently diagnosed with vasculitis or a change of their treatment due to a relapse (inception cohort). The GeVas registry allows long-term follow-up of a substantial cohort of vasculitis patients in a multicenter setting.

Objectives

To present the first data on the follow-up of newly diagnosed and relapsing GPA enrolled in the GeVas registry.

Methods

GeVas is a prospective, web-based, multicenter, clinician-driven registry for the documentation of organ manifestations, damage, long-term outcomes, and therapy regimens in various types of vasculitis. Recruitment started in June 2019. By January 2022, 17 centers in Germany were initiated and started enrolling patients. Meanwhile, more than 350 patients have been documented in the registry. Sites in Austria and the German-speaking cantons of Switzerland will be integrated soon [3].

Results

By mid-October 2021, the participating centers included 113 patients with GPA. The majority of patients were PR3-ANCA positive and affected by general symptoms, ENT, lung, renal, and neurological involvement. Patients commonly received cyclophosphamide or rituximab in combination with glucocorticoids for the induction of remission. Fewer patients received methotrexate or other immunosuppressants. Patient characteristics and therapy are summarized in Table 1.

Table 1.

Patient characteristics (n = 113). *Unless otherwise specified.

CategoryFeaturen (%)*AgeAge (years); median [range]60 [51 - 70]GenderMale61 (54.0)Female52 (46.0)Reason for inclusion in the registryNewly diagnosed vasculitis57 (51.4)Relapse56 (49.6)ANCA statusPR3-ANCA99 (87.6)MPO-ANCA4 (3.6)ANCA negative9 (7.9)Organ manifestationGeneral symptoms86 (76.1)ENT69 (61.1)Lung/chest66 (58.4)Renal35 (31.0)Cardiovascular7 (6.2)GI3 (2.7)Neurological27 (23.9)TherapyGlucocorticoid102 (90.3)Rituximab56 (49.6)Cyclophosphamide37 (32.7)Methotrexate and other immunosuppressants, respectively26 (23.0) and 19 (16.8), respectively

Conclusion

Here, we present the first interim analysis of the GeVas registry. Clinical manifestations of GPA reported herein show less frequent renal involvement in comparison with a recent report from another European registry (POLVAS) and an UK study [4, 5]. This is potentially related to the predominance of recruiting rheumatology centers thus far. By contrast, respiratory tract involvement is more frequent and PR3-ANCA less common in Japan [5]. Further data are prospectively documented and a follow up analysis is in progress.

References

[1]Hellmich B, et al. New insights into the epidemiology of ANCA-associated vasculitides in Germany: results from a claims data study. Rheumatology 2021;60:4868-73.

[2]Kitching AR, et al. ANCA-associated vasculitis. Nat Rev Dis Primers 2020;6:71.

[3]Iking-Konert C, et al. The Joint Vasculitis Registry in German-speaking countries (GeVas) – a prospective, multicenter registry for the follow-up of long-term outcomes in vasculitis. BMC Rheumatol 2021;5:40.

[4]Wójcik K, et al. Clinical characteristics of Polish patients with ANCA-asscoiated vasculitides – retrospective analysis of POLVAS registry. Clin Rheumatol 2019;38:2553-63.

[5]Furuta S, et al. Comparison of the phenotype and outcome of granulomatosis with polyangiitis between UK and Japanese cohorts. J Rheumatol 2017;44:216-22.

Acknowledgements

GeVas was supported by unrestricted grants by: DGRh, John Grube Foundation, Vifor and Roche Pharma

Disclosure of Interests

Sabrina Arnold: None declared, Pia Wallmeier: None declared, Fabian Schubach: None declared, Gabriele Ihorst: None declared, Peer Aries: None declared, Raoul Bergner Consultant of: VIFOR, Jan Philip Bremer: None declared, Norman Görl: None declared, Bernhard Hellmich: None declared, Jörg Henes: None declared, Bimba Hoyer: None declared, Antje Kangowski: None declared, Ina Kötter: None declared, Claudia Metzler: None declared, Ulf Müller-Ladner: None declared, Matthias Schaier: None declared, Ulf Schönermark: None declared, Jens Thiel: None declared, Leonore Unger: None declared, Nils Venhoff Speakers bureau: Roche and Vifor: speaker honoraries, Consultant of: Roche and Vifor: advisory boards, Grant/research support from: John-Grube Research Award 2021, Julia Weinmann-Menke: None declared, Jana Petersen: None declared, Christof Iking-Konert Speakers bureau: Lecture fees from: Chugai, GSK, Roche, and Vifor, Consultant of: Consulting fees from: Chugai, GSK, Roche, and Vifor, Grant/research support from: Research grants for GeVas: Roche, Vifor, DGRh, John Grube Foundation, Peter Lamprecht Speakers bureau: Chugai, GSK, Roche, and Vifor, Consultant of: Chugai, GSK, Roche, and Vifor, Grant/research support from: DGRh, John Grube Foundation, Roche, and Vifor

POS0800 THE JOINT VASCULITIS REGISTRY IN GERMAN-SPEAKING COUNTRIES (GeVas) – SUBGROUP ANALYSIS OF 131 GCA-PATIENTS REFERENCES:

Background

The most frequent form of vasculitis in elderly people is giant cell arteritis (GCA) with an annual incidence rate less than 10 per 100,000 persons over the age of 50. Like most vasculitides, GCA is characterized by chronicity and relapses, leading to significant overall morbidity and higher mortality in a subset of patients with aortic involvement and dissection. Most studies carried out so far have been retrospective, used monocentric study designs and small patient cohorts. Therefore, the Joint Vasculitis Registry in German-speaking countries (GeVas) has been established to record patients, who have been recently diagnosed with vasculitis or who have changed their treatment due to a relapse (inception cohort). The GeVas-Registry allows a long-term follow-up of a substantial cohort of vasculitis patients in a prospective and multicenter manner.

Objectives

To describe the subgroup of GCA and its characteristics within the GeVas registry.

Methods

GeVas is a prospective, web-based, multicenter, clinician-driven registry for the documentation of organ manifestations, damage, long-term outcomes, and therapy regimens in various types of vasculitis. Recruitment started in June 2019. By January 2022, 17 centers in Germany were initiated and have begun enrolling patients. Meanwhile, more than 350 patients have been documented in the registry. Sites in Austria and the German-speaking cantons of Switzerland will be integrated soon (1).

Results

By mid-October 2021, the participating centers recruited 131 GCA patients into the registry. 21.7% of patients (n=28) were enrolled in the registry due to relapse, and 78.3% (n=101) due to a first-time diagnosis. In accordance with long-standing epidemiology data, the majority of patients (67,2%), were female (n=88), and 32.8% (n=43) were male. Mean age was 74 years (max. 92y, min. 52y). The most frequently recorded organ manifestations in GCA patients addressed cranial and ophthalmic symptoms, and the cardiovascular system. However, vascular lung/chest involvement was also observed in 3% of cases (n=4). Out of the 131 patients, 97.7% (n=128) received immunosuppressive therapy, three refused to take any medication. An equal number of patients were treated with glucocorticoid (GC) therapy. While about two equal parts were treated by stable long term oral GC therapy (47,7%, n=62) or by i.v. pulse therapy followed by tapering (49,2%, n=64), only about 2.3% (n=3) were treated by oral GC therapy with intermittent i.v. pulses. 48.5% (n=63) of patients received tocilizumab as additional immunosuppressive therapy, 19.2% (n=25) methotrexate, and 18.5% (n=24) cyclophosphamide i.v. pulses.

Conclusion

In June 2019, we successfully established the prospective multicenter vasculitis registry GeVAS. It describes the first systematically recorded prospective GCA cohort in German-speaking countries. Its characteristics correspond to those that can be expected from the literature, with some unexpected finding e.g. the high proportion of patients treated with cyclosphosphamid, probably reflecting a sicker patient population with e.g. aortic or central nervous involvement. After 2.5 years of follow-up documentation, the first long-term results will be systematically evaluated and interpreted. The newly acquired data on disease manifestation, diagnostics and therapy regimens will provide important insights into the treatment of GCA patients in Germany and may generate further research goals.

References

Trial registration: German Clinical Trials Register (Deutsches Register Klinischer Studien): DRKS00011866. Registered 10 May 2019. 3

[1]C Iking-Konert; P Wallmeier; S Arnold; S Adler; K de Groot; B Hellmich; B Hoyer; K Holl-Ulrich; Ihorst; M Kaufmann; I Kötter; U Müller-Ladner; T Magnus; J. Rech; H. Schulze-Koops; N. Venhoff; T. Wiech; P. Villiger; F. Schubach; P. Lamprecht. The Joint Vasculitis Registry in German-speaking countries (GeVas) – a prospective, multicenter registry for the follow-up of long-term outcomes in vasculitis. BMC Rheumatol. 2021 Jul 31;5(1):40. doi: 10.1186/s41927-021-00206-2.

Acknowledgements

GeVas was supported by unrestricted grants by: DGRh, John Grube Foundation, Vifor and Roche Pharma

Disclosure of Interests

Pia Wallmeier: None declared, Sabrina Arnold: None declared, Fabian Schubach: None declared, Gabriele Ihorst: None declared, Peer Aries: None declared, Raoul Bergner Consultant of: Advisory Board VIFOR, Grant/research support from: John-Grube Research Award 2021, Jan Philip Bremer: None declared, Norman Görl: None declared, Bernhard Hellmich: None declared, Jörg Henes: None declared, Bimba Hoyer: None declared, Antje Kangowski: None declared, Ina Kötter: None declared, Tim Magnus: None declared, Claudia Metzler: None declared, Ulf Müller-Ladner: None declared, Matthias Schaier: None declared, Ulf Schönermark: None declared, Jens Thiel: None declared, Leonore Unger: None declared, Nils Venhoff Speakers bureau: Roche and Vifor, Consultant of: Roche and Vifor, Grant/research support from: John-Grube Research Award 2021, Julia Weinmann-Menke: None declared, Jana Petersen: None declared, Peter Lamprecht Speakers bureau: Lecture fees from: Chugai, GSK, Roche, Consultant of: Consulting & lecture fees from: Chugai, GSK, Roche, and Vifor., Grant/research support from: Research grants for GeVas: DGRh, John Grube Foundation, Roche, and Vifor, Christof Iking-Konert Speakers bureau: lecture fees from: Chugai, GSK, Roche, and Vifor., Consultant of: Consulting fees from: Chugai, GSK, Roche, and Vifor., Grant/research support from: Research grants for GeVas: DGRh, John Grube Foundation, Roche, and Vifor;

Outcome and clinical course of EHEC O104 infection in hospitalized patients: A prospective single center study

Objectives

Shiga-toxin producing O157:H7 Entero Haemorrhagic E. coli [STEC/EHEC] are the most common cause of Haemolytic Uraemic Syndrome [HUS] related to infectious haemorrhagic colitis. Nearly all recommendations on long term treatment of EHEC infections refer to this strain. The 2011 outbreak in Northern Europe was the first of this dimension to be caused by the serotype O104:H4. We report on the 3.5 year follow up of 61 patients diagnosed with symptomatic EHEC O104:H4 infection in spring 2011.

Methods

Patients with EHEC O104 infection were followed in a monocentric, prospective observational study at four time points: 4, 12, 24 and 36 months. These data include the patients’ histories, clinical findings, and complications.

Results

Sixty-one patients suffering from EHEC O104:H4 associated enterocolitis participated in the study at the time of hospital discharge. The mean age of patients was 43 ± 2 years, 37 females and 24 males. 48 patients participated in follow up 1 [FU 1], 34 patients in follow up 2 [FU 2], 23 patients in follow up 3 [FU 3] and 18 patients in follow up 4 [FU 4]. Out of 61 patients discharged from the hospital and included in the study, 54 [84%] were examined at least at one additional follow up. Serum creatinine decreased significantly between discharge and FU 1 from 1.3 ± 0.1 mg/dl to 0.7 ± 0.1 mg/dl [p = 0.0045]. From FU 1 until FU 4, no further change in creatinine levels could be observed. The patients need of antihypertensive medications decreased significantly [p = 0.0005] between discharge and FU 1 after four months. From FU 1 until FU 3, 24 months later, no further significant change in antihypertensive treatment was observed.

Conclusions

Our findings suggest that patients free of pathological findings at time of discharge do not need a specific follow up. Patients with persistent health problems at hospital discharge should be clinically monitored over four months to evaluate chronic organ damage. Progressive or new emerging renal damage could not be observed over time in any patient.