What is AI and Why You Should Be Excited About It

In 1950, Alan Turing asked, “Can machines think?”. Fast-forward to 2010 and artificial intelligence can diagnose diseases, fly drones, translate between languages, recognise emotions, trade stocks and even beat humans at Chess and Go.

Artificial Intelligence (AI), in essence, is machines mimicking human intelligence. Now, that can be of two types:
1. One that's already here, narrow AI, where a computer performs some very specific task. Take for example, Apple's Siri or Netflix's recommender system.
2. The other, general AI, that remains science fiction for now. If you're thinking of Jarvis in “Iron Man" or R2-D2 in “Star Wars”, you're quite right.

An application of AI is Machine Learning, where the computer automatically improves at performing a task, with more experience. Deep Learning is a subset of machine learning that is more intensive; uses more data and more complex algorithms.

Now, as a community of medicos, why should we bother about tech at all? Well, the future of healthcare looks increasingly facilitated by technology. The aim is to shift from "treating illness" to "sustaining wellness"; to have have a more proactive, rather than a reactive, model of care delivery. AI will help redesign our services and better utilise our resources. The goal isn't to replace what humans do, but instead augment it.

Here are a few ways AI can potentially help medicine:
1. Image recognition and diagnostic radiography, eg: Qure.ai, and Stanford's CheXpert system 
2. Preliminary diagnoses, eg: Babylon Health, and DeepMind's Streams application 
3. Virtual nursing assistants, eg: Care Angel's virtual nurse assistant
4. Clinical trials participant identifier, eg: deep6.ai
5. Computer-assisted robotic surgery, eg: Heartlander miniature robot
6. 3D mapping and printing, eg: 3D printed heart stents
7. Administrative workflow assistance, eg: IBM Watson 
8. Fraud detection and Cybersecurity, eg: H2O.ai 

[The list is by no means exhaustive. I implore you to know more about the examples mentioned by simply copy-pasting them into Google search.]

To conclude, the future of healthcare looks exciting and will be far more collaborative than it is today, working in alliance with AI, data science, statistics, engineering, and genomics. The ultimate objective is always to improve quality of treatment and patient outcomes.

Author's note: If, as a med student or a doctor, you're interested in kickstarting your own career towards AI and healthcare, please let me know in the comments. I will appropriately refer you to the relevant resources. To give you a brief background, I’ve worked with data scientists on seven medicine related portfolio projects, utilising machine and deep learning algorithms. I worked as a clinician and a programmer (have professional working proficiency in Python). Here’s my top 3:
1. Breast Cancer Detection Using Python & Machine Learning, with a model accuracy of 95% using artificial neural networks and support vector machine, on Wisconsin diagnostic data set
2. Identifying Skin Lesions Using Python & Deep Learning, with a model accuracy of 79% using convolution neural networks, on Cornell HAMNIST-10000 data set
3. Determining the Efficacy of Corrective Spinal Surgery in Childhood Kyphosis Using Python & Machine Learning, with a model accuracy of 88% using decision trees and random forest classifier on a Kaggle datset 

Thank you for reading.

- Ashish Singh

COVID-19: Whose Virus Is It Anyway? Possible origins of SARS-CoV-2

It's only reasonable you may want to know about the origins of the COVID-19 pandemic. After all, our lives have been affected, one way or the other. But was it the bat? Was it the pangolin? Or was it a lab experiment gone wrong? Let's look at the two most definitive evidence we have at hand: virus genomics and structure.

Evidence #1

The receptor binding domain (RBD) in the spike protein is the most variable part of the coronavirus family genome. SARS-CoV-2 seems to have an RBD that binds with high affinity to ACE2 from humans, and other species with high receptor homology. This RBD has six key amino acid residues.

Evidence #2

The second notable feature of SARS-CoV-2 is a polybasic cleavage site at the junction of S1 and S2, the two subunits of the spike. This allows effective cleavage by furin and other proteases and has a role in determining viral infectivity and host range. Insertion of proline to this site and subsequent addition of O-linked glycans are unique to SARS-CoV-2.

Keeping these in mind, we have:

Theory #1
Natural selection in animal before zoonotic transfer

As many early cases of COVID-19 were linked to the Huanan market in Wuhan, it is possible that an animal source was present at this location.

Given the similarity of SARS-CoV-2 to bat SARS-CoV-like coronaviruses, it is likely that bats serve as reservoir hosts for its progenitor. This "bat virus" or more formally, RaTG13 is nearly 96% identical to SARS-CoV-2. Its spike diverges in the RBD, which suggests that it may not bind efficiently to human ACE2. 

Malayan pangolins illegally imported into Guangdong province contain coronaviruses similar to SARS-CoV-2. Some "pangolin coronavirus" exhibit strong similarity to SARS-CoV-2 in the RBD, including all six key RBD residues. This clearly shows that the SARS-CoV-2 spike protein optimised for binding to human-like ACE2 is the result of natural selection.

Neither the bat nor the pangolin coronavirus, however, has polybasic cleavage sites. This means, no animal coronavirus has been identified that is sufficiently similar to be the direct progenitor of SARS-CoV-2. That said, the diversity of coronaviruses in bats and other species is massively undersampled. Mutations, insertions and deletions can occur near the S1–S2 junction of coronaviruses, which shows that the polybasic cleavage site can arise by a natural evolutionary process. This perfectly sets us up for our next theory.

Theory #2
Natural selection in human after zoonotic transfer

It is possible that a progenitor of SARS-CoV-2 jumped into humans to acquire the genomic features described above through adaptation, during undetected human-to-human transmission. Once acquired, these adaptations would enable the pandemic to take off.

All SARS-CoV-2 genomes sequenced so far have the genomic features described above and are thus derived from a common ancestor that had them too. The "pangolin coronavirus" has an RBD very similar to that of SARS-CoV-2, by the process of natural selection. From this, we can infer the same happened with the virus that jumped to humans. So we can say, with some degree of confidence, the insertion of polybasic cleavage site occured during human-to-human transmission.

From what we know the first case of COVID-19 has been traced back to November 2019. This presumes a period of unrecognised human-to-human transmission, between the initial zoonotic event and the acquisition of the polybasic cleavage site.

Theory #3
Lab experiment gone wrong

Basic research involving passage of bat SARS-CoV-like coronaviruses in cell culture and animal models has been ongoing for many years in biosafety level 2 laboratories across the world, and there are documented instances of laboratory escapes of SARS-CoV. In theory, it is possible that SARS-CoV-2 acquired RBD mutations during adaptation to passage in cell culture.

Having said that, the "pangolin coronavirus" with nearly identical RBDs, provides a much stronger explanation of how SARS-CoV-2 acquired these via recombination or mutation. The high-affinity binding of the SARS-CoV-2 spike protein to human ACE2 is most likely the result of natural selection on a human or human-like ACE2.

The acquisition of both the polybasic cleavage site and predicted O-linked glycans also argues against culture-based scenarios. New polygenic cleavage sites have only been observed after prolonged in-vivo passage whereas generating O-linked glycans likely involves an immune system.

Furthermore, if genetic manipulation had been performed, one of the several reverse-genetic systems available for coronaviruses would probably have been used. However, the genetic data irrefutably show that SARS-CoV-2 is not derived from any previously used virus backbone.

These are strong arguments that SARS-CoV-2 is not the product of purposeful manipulation.

Conclusion
Theory #2 seems most likely, given the information currently available, but more scientific data could swing the balance of evidence to favour one hypothesis over another. What's important is to further study the possible origins, not just for understanding the current zoonotic pandemic but also to prevent the potential future ones.

References
1. 'The proximal origin of SARS-CoV-2' by Andersen et al: www.nature.com/articles/s41591-020-0820-9
2. 'A pneumonia outbreak associated with a new coronavirus of probable bat origin' by Zhou et al: www.nature.com/articles/s41586-020-2012-7
3. 'A new coronavirus associated with human respiratory disease in China' by Wu et al: www.nature.com/articles/s41586-020-2008-3

- Ashish Singh

Coronary artery anatomy mnemonic and video for visualization

Let's learn about the coronary artery anatomy today (and never forget it!)

Watch the video. Text and images below.

Coronary artery dominance and EKG changes

Hello, hello!

Coronary arterial dominance is defined by the vessel which gives rise to the posterior descending artery (PDA).

Funnel Plot

-also called as Begg’s plot
-type of scatter plot
-used to examine biases in meta-analyses

An ideal funnel plot is symmetric.
If no biases, 95% of studies lie within the triangle.

Diabetic Ketoacidosis in just 5 minutes

Video by Drashtant 

Thioamides in pregnancy

Hello

Propylthiouracil is a pro. It always comes first (used in first trimester of pregnancy).
Methimazole causes Malformations in the embryo (teratogenic).

There are two M's in MethiMazole. This drug is used in second (and third trimester of pregnancy).
Propylthiouracil piles up, causing liver toxicity, thus limiting its use.

Hope it helps
- Jaskunwar Singh

Importance of Ischial spine

Following are the important points of the ishial spine :-

Mnemonic SID BPL

1) Station of fetal head is calculated with respect to Ischial spine.

2) Internal rotation of fetal head occurs at this level.

3) Deep transverse arrest occurs at this level.

4) It is site for giving pudendal block.

5) Place at which ring pessory inserted.

6) Levator ani muscle is attached here.

PS : Question which was asked in central institute examination (I felt it should be mentioed here)

Which ligament is felt while giving pudendal block?

Sacrospinal ligament.

Clinical pearl : TNF-alpha therapy

Hello

In case of granulomatous diseases, macrophages activated by Th1 cells lead to increased levels of TNF-alpha. Now, TNF-alpha induces and maintains granuloma formation. Basic, right?

So we give anti-TNF drugs (adalimumab, infliximab, etc.). However, they cause the granuloma to break down, thus leading to disseminated disease.

Bottom line - Always remember to check for the presence of latent TB before starting anti-TNF therapy.

That's all
- Jaskunwar Singh

COVID-19: effects on reproduction

Hello

In this post, I will be talking about effects of SARS-CoV-2 on the male reproductive system, as evidenced from a recent study.

COVID-19: Whatsapp group

I created a COVID-19 Whatsapp group to strictly discuss the medical aspect of the disease, the latest research/community practices. Email me if interested: medicowesome@gmail.com

-IkaN

COVID-19: Neurological manifestations

Since the Chinese health authorities confirmed the first case of novel coronavirus infection, almost all of the clinical focus has been on the viral's prodromal symptoms and severe life-threatening adverse effects such as ARDS. However, neurologists all over the world have been reporting the neurological manifestations of COVID-19 such as, ataxia, encephalopathy, myelitis among others. One neurological symptom in particular received inordinate attention, anosmia, even though it barely has any diagnostic relevance. It is safe to say that the neurological deficits are ongoing in this pandemic without getting noticed appropriately. However, since we are in the early phases of understanding the clinical conundrum of the COVID-19, such relative blindness is expected.

How does SARS-CoV-2 enter the CNS?

Two pathways have been postulated:
1. Through the cribriform plate
2. Systemic circulatory dissemination after infecting the lungs.

Reported neurological manifestations:

1. Anosmia - Can be explained by the proximity of the olfactory bulb to the cribriform plate
2. Hypoguesia, dysguesia
3. Headache, malaise
4. Unstable walking or ataxia, dizziness
These four can occur in the early phase of the disease.

5. Cerebral hemorrhage - This has been hypothesized to be due to decrease in expression and function of ACE2 proteins, especially in hypertensive patients in whom the expression of ACE2 is already low. Given that ACE2 signaling lowers BP, lack of ACE2 function would lead to higher BP which might precipitate cerebral hemorrhage.
6. Cerebral infarction (acute cerebrovascular disease causing stroke)
7. Ondine's curse - The central respiratory centres lose their function, which consequently impairs involuntary respiration severely.
8. Acute encephalopathy - headache, altered mental status, convulsions.
9. Myopathy

Interestingly, the CSF in the patients were normal, which implies that COVID-19 does not cross the blood brain barrier and hence cannot cause meningitis or encephalitis. We should keep in mind that the neurological manifestations could be secondary to hypoxia, respiratory or metabolic acidosis and other complications of the COVID-19 infection.

Thank you!

-Vinayak

References:

1. Necrotizing Encephalopathy: CT and MRI Features

https://pubs.rsna.org/doi/10.1148/radiol.2020201187

2. Neurological Complications of Coronavirus Disease (COVID-19): Encephalopathy

https://www.cureus.com/articles/29414-neurological-complications-of-coronavirus-disease-covid-19-encephalopathy

3. Neurological Manifestations of Hospitalized Patients with COVID-19 in Wuhan, China: a retrospective case series study

https://www.medrxiv.org/content/10.1101/2020.02.22.20026500v1

COVID-19: Lymphopenia and pneumonia

Hello everyone!

In the context of COVID-19, we will talk about two specific terms: Lymphopenia and Pneumonia.

COVID-19 Pneumonia
We mention "pneumonia" when there is an acute inflammation of the lungs following an infection. Pneumonia is one of the common features in infected patients with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). This pneumonia has various clinical and radiological characteristics depending on the stage of the disease. It evolves rapidly, even in asymptomatic patients from local unilateral to diffuse bilateral ground-grass opacities which progress within 1-3 weeks to consolidation or co-exists with. A retrospective study at Wuhan describes radiological findings from 81 patients with COVID-19 pneumonia. The predominant pattern of abnormality observed was bilateral (79%), peripheral (54%), ill-defined (81%) and ground-glass opacification (65%), mainly involving the right lower lobes. [1]

Instructions for new authors: Images, plagiarism, and grammar

Hello awesome authors,

I thought of writing a small guide on things to be mindful when posting images or writing new blogs.

COVID-19: Use of masks

Hi everyone!

We used the WHO guidelines to write the pdf and uploaded it over here

COVID 19: How to limit the spread?

COVID19 spreads primarily through droplets of saliva or discharge from the nose when an infected patient coughs or sneezes (we should so cough or sneeze into a tissue or flexed elbow). The SARS-CoV-2 can also be carried, that's why the handwashing is so important.

We use other means of prevention to limit the spreading, for example, masks and negative pressure rooms. Let us see how it is done.

COVID-19: Containment strategy by South Korea

Hello everyone!

In this post, we will discuss the manner with which South Korea managed to contain the virus rather successfully.

So let me help you catch up:-

Club foot: Age-wise Management Flowchart

Club foot is one that resembles a golf club. It is also called Congenital Talipes Equino Varus or CTEV.

Figure 1. Dennis-Brown Splint

How to cite articles

Hi everyone,

I wanted to write a quick post on how to cite references for Medicowesome Student Guest Authors (MSGAites!). Medicowesome is not a peer-reviewed journal, we are just a website where we post mnemonics, study material, and cool facts. Recently, we've been writing about COVID-19. Because there has been so much fake news and miscommunication about the characteristics of this disease, we decided that all posts related to COVID-19 would have journal articles in literature as references.

There are many styles in which you can format references. You can read more about it in this paper by Kambhampati & Maini, 2019. [1] It is preferred that you use a particular formatting style for all the references in your article. Simply adding links is not preferred because websites change their links all the time. The best way to ensure that your reader finds the article you're referencing is by using a proper reference format. A DOI is guaranteed never to change, so you can use it as a permanent link to any electronic article.

COVID-19: SARI treatment facility design

Hi everyone,

One of our guest authors, Tanay Saxena, recently completed a course on Severe Acute Respiratory Infections Treatment Centre. He compiled a very thorough set of notes during the course based on the WHO Severe Acute Respiratory Infections Treatment Centre practical manual that has been developed for the COVID-19 pandemic.

COVID-19: Trained immunity from BCG vaccine

Would BCG vaccination really help in immunizing up against SARS-CoV-2?

Let's dig in. 

BCG is a live-attenuated strain derived from an isolate of Mycobacterium bovis used widely across the world as a vaccine for tuberculosis (TB). But that's not all, BCG vaccination is a potential goldmine against so many diseases.