Gabriela Czanner

Home

Research

Research Group

Teaching

Software

Meetings

Consultations

Memberships

 

News

 

 

 

Liverpool Group for Statistical Spatio-Temporal Modelling of Medical Images

Research

The primary aim of research in the ophthalmic statistics group is the development of the novel methodology to design and analyses ophthalmic studies. Currently we are addressing a number of key challenges in inference and prediction from observational studies:

1.       Spatial statistical models clinical images toward understanding of ophthalmic diseases.

Figure 4

Manual grading of lesions in retinal images is relevant to clinical management and clinical trials, but it is time-consuming and expensive. Furthermore, it collects only limited information - such as lesion size or frequency. The spatial distribution of lesions is ignored, even though it may contribute to the overall clinical assessment of disease severity, and correspond to microvascular and physiological topography. Capillary non-perfusion (CNP) lesions are central to the pathogenesis of major causes of vision loss. With clinicians Ian MacCormick and Simon Harding and imaging specialists Yalin Zheng, Silvester Czanner we proposed a novel method to analyse CNP using spatial statistical modelling. This quantifies the percentage of CNP-pixels in each of 48 sectors and then characterises the spatial distribution with goniometric functions. We applied our spatial approach to a set of images from patients with malarial retinopathy, and found it compares favourably with the raw percentage of CNP-pixels and also with manual grading. Furthermore, we were able to quantify a biological characteristic of macular CNP in malaria that had previously only been described subjectively: clustering at the temporal raphe. Microvascular location is likely to be biologically relevant to many diseases, and so our spatial approach may be applicable to a diverse range of pathological features in the retina and other organs. This work is now being extended in our group into spatio-temporal models.

 

2.       Longitudinal and spatio-temporal models for early detection, screening and monitoring of diseases. We recently developed methods of longitudinal discriminant analysis allow for classification of subjects into prespecified prognostic groups using longitudinal history of both continuous and discrete biomarkers (Hughes et al, 2016). The classification uses Bayesian estimates of the group membership probabilities for each prognostic group.

Currently, a PhD student in my group, Wenyue Zhu, is extending the model to account for longitudinally collected images, in collaboration with Prof Simon Harding, imaging specialists Yalin Zheng and statistician Ruwanthi Kolamunnage-Dona. My student Mohamed Barkat is looking into proposing dynamic monitoring for endovascular surgeries.

 

3.       Signal-to-noise ratio metrics for evaluation of the signal in the complex data and for feature extraction.

Fig. 1.

Neurons represent both signal and noise in binary electrical discharges termed action potentials. Hence, the standard signal-to-noise ratio (SNR) definition of signal amplitude squared and divided by the noise variance does not apply. We show that the SNR estimates a ratio of expected prediction errors. Using point process generalized linear models, we extend the standard definition to one appropriate for single neurons. In analyses of four neural systems, we show that single neuron SNRs range from −29 dB to −3 dB and that spiking history is often a more informative predictor of spiking propensity than the signal or stimulus activating the neuron. By generalizing the standard SNR metric, we make explicit the well-known fact that individual neurons are highly noisy information transmitters.

PhD student Padraic Walsh, in my group, has been developing a new SNR metric that can be used for the selection of best set of predictive biomarkers from clinical observational data, especially when data are coming from varying modalities (e.g. image, genetics, physiology).

 

4.       Human, statistical and machine learning for disease understanding from complex data. We look into comparison of human, statistical and machine approaches to complex datasets; validation of generalisability and application to problems from medicine such as glaucoma, keratoconus and gait analyses. This is supported by EPSRC GCRF (PI: Yalin Zheng), EPSRC Mathematics Healthcare Centre (PI: Ke Chen).

The findings of our research contributes to ophthalmic research in the following area: diabetes related diseases, age-related diseases and malaria related diseases.

 

Selected research papers of the group

1.         Barrera V, MacCormick IJC, Czanner G, et al., Neurovascular sequestration in paediatric P.falciparum malaria is visible clinically in the retina. eLife 2018;7:e32208 DOI: 10.7554/eLife.32208 (IF 7.7)

2.         MacCormick IJC, Zheng Y, Czanner S, Zhao Y, Diggle PJ, Harding SP and Czanner G. Spatial statistical modelling of capillary non-perfusion in the retina. Scientific Reports 7, 2017 https://www.nature.com/articles/s41598-017-16620-x

3.         Sahni J, Czanner G, Gutu T, Kate M Bennett K, Wuerger SM, Grierson I, Murray-Dunning C, Holland MN,  Harding SP, Safety and acceptability of an organic light emitting diode sleep mask as a potential therapy for retinal disease. Eye 2017; 31(1):97-106. PMCID: PMC5233944  (IF 2.1)

4.         Ho WMV, Hussain R, Czanner G, Heimann H, Damato BE. Porous vs non-porous orbital implants after enucleation for uveal melanoma: a randomized study. Ophthal Plast Reconstr Surg. 2016 [Epub ahead of print] PMID: 27861329 (IF 0.99)

5.         Silvester A, Neal T, Czanner G, Briggs M, Harding S, Kaye S. Adult bacterial conjunctivitis: resistance patterns over 12 years in patients attending a large primary eye care centre in the UK. BMJ Open Ophthalmol, 1(1) DOI: 10.1136/bmjophth-2016-000006  (IF 3.04)

6.         Aslam TM, Tahir HJ, Parry NRA, Murray IJ, Kwak K, Heyes R, Saleh MM, Czanner G, Ashworth J. Automated measurement of visual acuity in pediatric ophthalmic patients using principles of game design and tablet computers. Am J Ophthalmol. 2016(170):223-227. (IF 3.8)

7.         Hughes D, Komarek A, Czanner G, Garcia-Finana M. Dynamic longitudinal discriminant analysis using multiple longitudinal markers. Stat Methods Med Res. First published date: October 2016. DOI: 10.1177/0962280216674496  (IF 4.6)

8.         Knox PK, MacCormick IJC, Mbale E, Malewa M, Czanner G, Harding SP. Longitudinal visuomotor development in a malaria endemic area: cerebral malaria and beyond. PLoS One . 2016;11(10):e0164885. doi:10.1371/journal.pone.0164885. PMCID: PMC5072745  (IF 3.2)

9.         Chan YK, Czanner G, Shum HC, Williams RL, Cheung N, Wong D. Towards better characterisation and quantification of emulsification of silicone oil in vitro. Acta Ophthalmol 2016. [Epub ahead of print]  DOI: 10.1111/aos.13258  (IF 3.0)

10.     Vallabh NA, Sahni JN, Parkes CK, Czanner G, Heimann H, Damato B. Near infrared reflectance and autofluorescence imaging characteristics of choroidal nevi. Eye 2016. DOI: 10.1038/eye.2016.183. (IF 2.1)

11.     Aslam TM, Parry NRA, Murray IJ, Salleh M, Dal Col C, Mirza N, Czanner G, Tahir HJ. Development and testing of an automated computer tablet-based method for self-testing of high and low contrast near visual acuity in ophthalmic patients. Graefes Arch Clin Exp Ophthalmol. 2016; 254(5):891-899. (IF 1.9)

12.     Mittal R, Araujo I, Czanner G, Coupland S. Perforin expression in eyelid sebaceous carcinomas: a useful and specific immunomarker for the differential diagnosis of eyelid carcinomas. Acta Ophthalmol 2016; 94(5):e325-e330. (IF 3.0)

13.     Cox T, Czanner G. A practical divergence measure for survival distributions that can be estimated from Kaplan-Meier curves. Stat Med 2016; 35(14):2406-2421. (IF 2.0)

14.     Tidbury L, Czanner G, Newsham D. Fiat Lux: The effect of illuminance on visual and stereo acuity testing, Graefes Arch Clin Exp Ophthalmol 2016; 254(6):1091-1097. (IF 1.9)

15.     Balaskas K, Tiew S, Czanner G, Tan AL, Ashworth J, Biswas S, Aslam T. The novel evidenced assessment of tortuosity system: interobserver reliability and agreement with clinical assessment. Acta Ophthalmol Online 21 Dec 2015. DOI: 10.1111/aos.12907.  (IF 3.0)

16.     MacCormick IJC, Czanner G, Faragher B, Developing retinal biomarkers of neurological disease – an analytical perspective. Biomark. Med. 2015; 9(7): 691–701. (IF 2.2)

17.     Boonarpha N, Zheng Y, Stangos AN, Lu H, Raj A, Czanner G, Harding SP, Nair-Sahni J, Standardisation of choroidal thickness measurements using enhanced depth imaging optical coherence tomography, Int J Ophthalmol 2015 18;8(3):484-91. (IF 0.9)

18.     Czanner G, Sarma SV, Ba D, Eden UT, Wu W, Eskandar E, Lim HH, Temereanca S, Suzuki WA and Brown EN. Measuring the signal-to-noise ratio of a neuron. Proc Natl Acad Sci U S A. 2015 Jun 9;112(23):7141-6. (IF 9.4)

19.     Stewart RMK, Sheridan C.M, Hiscott PS, Czanner G, Kaye SB. Human conjunctival stem cells are predominantly located in the medial canthal and inferior forniceal areas. Invest Ophthalmol Vis Sci. 2015 Feb 26;56(3):2021-30. (IF 3.5)

20.     Spiteri N, Siradas G, Czanner G, Batterbury M, Kaye SB. Assessing the quality of ophthalmic anaesthesia. J Clin Anesth 2015 June; 27(4):285-9. (IF 1.3)

21.     Cumberland PM, Czanner G, Bunce C, Dore CJ, Freemantle N, Garcia-Finana M. Ophthalmic statistics note: the perils of dichotomizing continuous variables. Br J Ophthalmol 2014;98: 841-843. (IF 3.0)

22.     Raj A, Sahni JN, Campa C, Czanner G, Hagan R, Harding SP. Macular function: better than visual acuity for assessing diabetic maculopathy? Eur J Ophthalmol  2013; 23(3): 459-460. (IF 1.0)

23.     Green J, Czanner G, Reeves G , Watson J, Wise L, Roddam A, Beral V.  Menopausal hormone therapy and risk of gastrointestinal cancer: prospective study and meta-analyses. Int J Cancer 2012;130(10): 2387-96. (IF 5.5)

24.     Sarma S, Nguyen D, Czanner G, Wirth S, Suzuki W, Wilson MA, Brown E. Error analysis for point process adaptive filters: A bootstrap approach. Neural Comput 2011;23(11): 2731-45. (IF1.6)

25.     Petrasova A, Czanner S, Czanner G, Farrer J, Chalmers A, Wolke D. The influence of dynamic scene changes in positioning and manipulating VR tasks. J Appl Math 2010;3(1): 259-268. (IF 1.5)

26.     Green J, Czanner G, Reeves G, Watson J, Wise L, Beral V. Oral bisphosphonates and risk of cancer of the oesophagus, stomach and colorectum: nested case-control study. Br Med J 2010;341(7772): 545-552. (IF 10.0)

27.     Czanner G, Eden UT, Wirth S, Yanike M, Suzuki WA, Brown EN. Analysis of between-trial and within-trial neural spiking dynamics. J Neurophysiol 2008;99(5): 2672-2693.  (IF 4.6)

28.     Czanner G, Grün S, Iyengar S. Theory of the snowflake plot and its relations to higher-order analysis methods. Neural Comput 2005;17(7): 1456-1479 (IF 1.6)

29.      Krimer LS, Zaitsev AV, Czanner G, Kroner S, Gonzalez-Burgos G, Povysheva NV, Iyengar S, Barrionuevo G, Lewis DA. Cluster analysis-based physiological classification and morphological properties of inhibitory neurons in layers 2-3 of monkey dorsolateral prefrontal cortex. J Neurophysiol 2005;94(5): 3009-3022 (IF 2.9)

 

 

Members

Dr Gabriela Czanner

Lecturer in Ophthalmic Biostatistics and the leader of the group.

Ms Wenyue Zhu

PhD student at Institute of Ageing and Chronic Diseases, working on spatio-temporal modelling of ophthalmic images. (Since 2018)

Mr Joshua Bridge

Mres student working on temporal modelling (2018)

Dr Joe Butler (former member)

Dr Butler is a Biostatistics PDRA working on statistical models for prediction of conversion to wet age-related macular degeneration (project EDiMAD, Sept 2016-Feb 2017).

Dr Ian MacCormick

Dr MacCormick is associated with Department of Eye and Vision Science. He collaborates with Gabriela on statistical methods for evaluation of surrogate markers and on spatial analysis of retinal images.

Mrs Riham El Saeiti 

Ms Riham El Saeiti is a PhD student at Department of Biostatistics. She is working on discriminant function analysis methods, mixed effect models and applications in medicine. Co-advised with Marta Garcia-Finana. Start 2013.

Ms Nazatulshima Hassan

Ms Nazatulshima Hassan is a PhD student at Department of Biostatistics. She is working on variable selection for discriminant analysis for nominal data and logistic regression: application to SNP data. Co-advised with Marta Garcia-Finana and Andrea Jorgensen. Start 2014.

Mr Padraic Walsh (former member)

Mr Padraic Walsh is a Phd student at Department of Biostatistics. He is working on variable selection for discrimination for complex health care data, signal-to-noise ratio, quadratic discriminant analysis: application to diabetic retinopathy and imaging. Co-advised with Simon Harding and Marta Garcia-Finana. Start Oct 2013.

Ms Henal Desai (former member)

Ms Henal Desai is a MRes Clinical Sciences student at IACD. She is working on developing a statistical model for predicting technical complications following endovascular surgery for aortic aneurysms, together with Gabriela, Mr Iain Roy (Vascular Research Fellow) and Professor S R Vallabhaneni (Professor of Vascular Surgery and Consultant Vascular & Endovascular Surgeon) from Liverpool Vascular & Endovascular Service. March-July 2016 Henal won the IACD MREs Best Poster Price for her project!

Alumni

 

Chris Bell, Manoharan Kumararaj and Thomas Hughes

Role of HbA1c, Cholesterol and Blood pressure in progression to sight threatening diabetic retinopathy. Clinical Research Students, 3rd year, Sept 2015-April 2016.

Ashwin Balu (3rd year clinical student)

Role of HbA1c in diabetic retinopathy. Sept 2014-April 2015.

Amanda Waleh-Astrom (Mres student)

Ordinal logistic regression and application in diabetic retinopathy. Sept-Dec 2014.

 

Contact: czanner@liverpool.ac.uk

Address: University of Liverpool, William Duncan Building, Room 287, Liverpool, L69 3GA

Last updated: 15 May 2018