Imagine a future where a diagnosis for a baffling illness isn’t a long, anxiety-ridden journey through countless specialists and tests, but a swift, accurate assessment, potentially even before you leave the primary care office. That future might be closer than we think, thanks to some truly groundbreaking work from Microsoft. Their new AI system, dubbed the AI Diagnostic Orchestrator (MAI-DxO), is showing an almost unbelievable ability to diagnose diseases, outperforming human generalist physicians by a margin that has everyone in the medical community talking.

This isn’t just about incremental improvements; we’re talking about a dramatic leap forward in AI in disease diagnosis. The implications are enormous, not just for patient outcomes but for the entire healthcare system, from costs to accessibility. It’s a development that’s making waves, and frankly, it should. Let’s dig into what this means and why it’s such a game-changer.

1. The Unveiling of MAI-DxO: A New Era in Diagnostics

Microsoft’s AI Diagnostic Orchestrator (MAI-DxO) isn’t just another flashy tech demo; it’s a meticulously developed system poised to shake up medical diagnostics. What makes it so powerful? It’s built on the cutting-edge foundation of OpenAI’s GPT-5 and GPT-5.5 models. For those not deep into AI, think of these as some of the most advanced large language models available, capable of understanding and processing complex information in ways that were science fiction just a few years ago. By integrating these models, Microsoft has created an AI that can not only sift through vast amounts of medical data but also reason and infer in a way that mimics, and in some cases surpasses, human clinical judgment.

The system’s name, “Diagnostic Orchestrator,” hints at its sophisticated functionality. It’s not just a pattern matcher; it’s designed to orchestrate the diagnostic process, considering various symptoms, patient history, test results, and differential diagnoses, much like a seasoned physician would, but with a speed and breadth of knowledge that no single human could ever achieve. This level of sophistication in AI in disease diagnosis represents a significant milestone, moving beyond simple data interpretation to complex clinical reasoning.

To truly appreciate the “orchestrator” aspect, consider how a human doctor approaches a complex case. They gather symptoms, review medical history, order tests, analyze results, and then synthesize all this information to form a differential diagnosis – a list of possible conditions. MAI-DxO mimics this process but at an accelerated pace and with access to a knowledge base far exceeding any individual physician’s. It can cross-reference symptoms with millions of documented cases, research papers, and clinical guidelines in seconds, identifying rare presentations or subtle connections that might take a human doctor hours or even days to uncover. This isn’t just about speed; it’s about the ability to connect disparate pieces of information with an unparalleled level of comprehensive analysis, moving beyond a simple checklist approach to a truly integrated diagnostic workflow.

2. Mind-Blowing Accuracy Rates: Doubling Human Performance

Now, let’s talk numbers, because this is where MAI-DxO truly shines and where the controversy, if you want to call it that, really begins. In rigorous tests involving intricate medical cases – the kinds that often stump even experienced doctors – MAI-DxO didn’t just perform well; it excelled. The system achieved a staggering 83.2% diagnostic accuracy. Compare that to practicing generalist physicians, who, even with access to reference materials, managed only 36.1% accuracy on the same cases. That’s more than double the accuracy. Think about that for a moment: an AI system more than twice as good at diagnosing complex diseases than human doctors using their usual tools.

And it wasn’t a fluke. In another set of tests focusing on emergency department admissions, where quick, accurate diagnoses are literally life-saving, MAI-DxO hit 80.3% accuracy, while human doctors achieved 46.8%. These figures, published in the esteemed peer-reviewed journal Science on October 22, 2026, aren’t just impressive; they’re transformative. They suggest that the potential for AI in disease diagnosis isn’t just theoretical; it’s already here, demonstrating capabilities that could redefine medical practice.

These accuracy rates aren’t just statistical curiosities; they represent a tangible impact on patient lives. A misdiagnosis or delayed diagnosis can lead to worsening conditions, inappropriate treatments, or even fatalities. By significantly improving diagnostic accuracy, especially in complex or emergency situations, MAI-DxO offers a lifeline. Consider rare diseases, which often go undiagnosed for years, leading to immense patient suffering and healthcare costs. An AI like MAI-DxO, with its vast knowledge and pattern recognition capabilities, could potentially identify these conditions much earlier, drastically improving patient prognoses. This level of precision in AI in disease diagnosis moves beyond simply assisting doctors; it positions AI as a critical partner in preventing diagnostic errors that currently plague healthcare systems worldwide.

3. Cost Reduction Potential: Streamlining Healthcare Expenses

Beyond accuracy, one of the most compelling aspects of MAI-DxO is its potential to dramatically reduce healthcare costs. Medical expenses are a huge burden for individuals and entire nations, and a significant portion of that cost comes from diagnostic processes – often, the ordering of numerous, sometimes unnecessary, tests. Microsoft’s research indicates that their AI can achieve accurate diagnoses at half, or even one-third, of the current cost, primarily by ordering fewer, more targeted tests.

How does it do this? By leveraging its vast knowledge base and sophisticated reasoning capabilities, MAI-DxO can more accurately pinpoint the most probable diagnoses earlier in the process. This means less trial-and-error, fewer redundant tests, and a more efficient allocation of resources. Imagine the financial relief for patients and healthcare systems if diagnostic pathways become significantly more streamlined and less expensive. This economic benefit alone makes advancements in AI in disease diagnosis incredibly attractive to policymakers and hospital administrators alike.

To put this into perspective, think about the current diagnostic merry-go-round many patients experience. You might visit a primary care doctor, get referred to a specialist, who then orders a battery of tests (blood work, imaging, biopsies), some of which might be redundant or ultimately unhelpful, simply because the initial diagnosis isn’t clear. Each of these steps incurs costs – consultation fees, lab charges, imaging expenses, and the patient’s time off work. MAI-DxO aims to cut through this by offering a high-probability diagnosis much earlier, suggesting only the most relevant and necessary tests to confirm it. This isn’t about cutting corners; it’s about optimizing the diagnostic journey based on an unparalleled analysis of symptoms and patient data, ensuring that every test ordered serves a direct, high-value purpose. The savings could be astronomical, redirecting resources from unnecessary diagnostics to actual treatment and preventative care, ultimately making healthcare more sustainable and accessible for everyone.

4. The Human Element: What Does This Mean for Doctors?

Naturally, when an AI system is shown to outperform human professionals, questions arise about the future of those professions. Does this mean AI will replace doctors? Not necessarily, and probably not entirely. While MAI-DxO’s diagnostic capabilities are clearly superior in these specific tests, medicine is far more than just diagnosis. It involves empathy, communication, surgical skill, hands-on care, and the nuanced understanding of a patient’s life context.

What’s more likely is that AI in disease diagnosis will become an indispensable tool, a super-powered assistant for physicians. Doctors could leverage MAI-DxO to quickly narrow down complex cases, confirm suspicions, or even flag rare conditions they might not immediately consider. This frees up human doctors to focus on the aspects of care that truly require a human touch – building rapport with patients, explaining complex prognoses, and providing emotional support. It’s about augmentation, not replacement, allowing doctors to practice at the very top of their game with an unparalleled diagnostic partner by their side. (See: NIH initiative on AI in healthcare.)

Consider the daily pressures on a general practitioner: a packed schedule, diverse patient complaints, and the constant need to stay updated on medical advancements across countless specialties. MAI-DxO can act as a tireless second opinion, a knowledge base that never forgets, and a rapid research assistant. Instead of spending precious minutes poring over medical texts or online databases for an obscure symptom, a doctor could feed patient data into MAI-DxO and receive a prioritized list of potential diagnoses and recommended tests, instantly. This doesn’t diminish the doctor’s role; it elevates it. They become more efficient, more accurate, and have more time to engage with patients on a human level, discussing lifestyle changes, emotional well-being, and treatment preferences, which are areas where AI simply cannot replicate the human experience. The integration of AI in disease diagnosis allows physicians to move beyond rote memorization and towards a more holistic, patient-centered approach, supported by cutting-edge technology.

5. Clinical Testing on the Horizon: Beth Israel Deaconess Medical Center

This isn’t just a lab experiment or a theoretical concept. Microsoft’s AI chief, Mustafa Suleyman, has confirmed that MAI-DxO is moving into real-world clinical testing. Within the next few months, Beth Israel Deaconess Medical Center in Boston is slated to begin trials. This is a critical next step. While lab results are impressive, the true test of any medical technology comes in the messy, unpredictable environment of a hospital or clinic.

Clinical trials will provide invaluable data on how MAI-DxO performs with actual patients, under the direct supervision of medical professionals. It will help refine the system, identify any unforeseen challenges, and build trust among the medical community. This move from research to practical application is what truly signifies the readiness of this advanced AI in disease diagnosis to start making a tangible difference in healthcare. For more context, see New Hepatitis B Breakthrough.

The transition to clinical trials at a prestigious institution like Beth Israel Deaconess Medical Center is a massive validation. It means the system has passed initial rigorous evaluations and is considered safe and promising enough to be tested on real patients, albeit under controlled conditions. These trials will not only assess accuracy in diverse patient populations – accounting for variations in age, ethnicity, medical history, and socioeconomic factors – but also evaluate user experience for physicians. How easily does it integrate into existing electronic health record (EHR) systems? Is the interface intuitive? Does it provide explanations for its diagnoses that doctors can understand and trust? The feedback from clinicians on the ground will be instrumental in refining MAI-DxO into a truly practical and indispensable tool. This phase is about moving beyond theoretical potential to proven, real-world utility for AI in disease diagnosis, addressing the practicalities of deployment in complex healthcare settings.

6. The Viral Narrative: AI vs. Human Doctors

It’s no surprise that this story has gone viral. The narrative of artificial intelligence outperforming human experts, especially in a field as sensitive and critical as medicine, is inherently captivating and emotionally charged. People are fascinated, intrigued, and perhaps a little unsettled by the idea of machines diagnosing better than doctors. It taps into both our hopes for advanced technology and our anxieties about automation.

This particular angle – AI in disease diagnosis achieving superior accuracy – resonates deeply because it touches on fundamental human concerns about health, life, and the role of human expertise. It’s a glimpse into a future that feels both thrilling and slightly unsettling, prompting widespread discussion and debate across social media, news outlets, and dinner tables. The emotional response is powerful, reflecting the profound implications of such a development.

The “AI vs. Human” narrative, while sensationalized, highlights a genuine tension. On one hand, there’s excitement about the potential for improved health outcomes and a more efficient healthcare system. On the other, there’s a natural human fear of being replaced, especially in professions that require extensive training and deep human connection. This public discourse is crucial, as it forces us to confront not just the technological capabilities but also the societal implications of such powerful AI tools. It pushes developers to consider ethical guardrails and ensures that the public’s concerns about safety, privacy, and accountability are addressed proactively. The viral nature of this story means that the conversation around AI in disease diagnosis isn’t confined to medical journals or tech conferences; it’s happening everywhere, shaping public perception and ultimately influencing the speed and manner of its adoption.

7. Monetization Potential: A Lucrative Niche for Healthcare AI

From a business perspective, the commercial potential of MAI-DxO and similar AI in disease diagnosis tools is immense. This technology falls squarely within the lucrative medical/healthcare and B2B SaaS (Software as a Service) niches. Hospitals, clinics, and even individual practitioners represent a massive market for advanced diagnostic tools that can improve patient outcomes and reduce operational costs. The demand for efficiency and accuracy in healthcare is constant, and AI offers a compelling solution.

This creates significant opportunities for commercial search intent around terms like “AI diagnostic tools,” “healthcare AI solutions,” and “medical technology investment.” Companies developing or integrating such AI systems stand to gain substantial market share. Furthermore, the ecosystem around this technology could foster affiliate opportunities related to health tech, AI in medicine, and even medical training programs that incorporate AI literacy.

The monetization strategy for a tool like MAI-DxO could take several forms. A subscription-based SaaS model for hospitals and large clinics is highly probable, offering tiered services based on usage volume or specific diagnostic modules. There’s also potential for specialized licensing agreements with pharmaceutical companies for drug discovery or with insurance providers for more accurate risk assessment and claims processing. The global health tech market is projected to reach trillions, and diagnostic AI is poised to capture a significant portion of that growth. Beyond direct sales, the data generated by these systems, anonymized and aggregated, could be invaluable for public health research, disease pattern tracking, and even policy-making, opening up further revenue streams through data analytics services. The sheer scale of the healthcare industry means that even a small percentage of market penetration for AI in disease diagnosis translates into billions of dollars, making it a highly attractive sector for investment and innovation.

8. Ethical Considerations and Implementation Challenges: Navigating the Future

While the benefits of MAI-DxO are clear, the path to widespread adoption is not without its hurdles. Ethical considerations, for instance, are paramount. Who is accountable if an AI makes a diagnostic error? How do we ensure equitable access to such advanced technology, preventing a two-tiered healthcare system? And what about data privacy and security, especially when dealing with highly sensitive patient information?

Beyond ethics, there are significant implementation challenges. Integrating complex AI systems into existing, often outdated, healthcare IT infrastructures is no small feat. Training medical professionals to effectively use and trust these tools will also be crucial. These are not insurmountable problems, but they require careful planning, robust regulatory frameworks, and a collaborative effort between tech developers, medical practitioners, policymakers, and ethicists. Addressing these challenges thoughtfully will be key to realizing the full promise of AI in disease diagnosis.

Elaborating on accountability, if MAI-DxO suggests a diagnosis that leads to an incorrect treatment and adverse patient outcomes, where does the legal and ethical responsibility lie? Is it with Microsoft, the hospital, the doctor who followed the AI’s recommendation, or a combination? Clear legal frameworks and robust malpractice guidelines will need to be established to navigate these complex scenarios. On data privacy, the AI needs access to vast amounts of patient data to learn and perform effectively. Ensuring this data is anonymized, secured against breaches, and used ethically, without bias, is a monumental task. The potential for algorithmic bias, where the AI might perform less accurately for certain demographic groups due to biases in its training data, also needs to be rigorously addressed. Overcoming these ethical and implementation hurdles will require not just technological prowess but also a deep understanding of human values and societal impact, ensuring that the integration of AI in disease diagnosis truly serves all of humanity.

9. The Broader Impact on Medical Education and Practice: A Shifting Paradigm

This leap in AI in disease diagnosis isn’t just about a new tool; it represents a shifting paradigm for medical education and the very practice of medicine. Future doctors won’t just need to master anatomy and pharmacology; they’ll need to be adept at interacting with AI systems, understanding their outputs, and knowing when to trust them and when to apply critical human judgment. The curriculum for medical schools will undoubtedly evolve to include AI literacy, data science, and perhaps even a focus on the unique human skills that AI cannot replicate.

For current practitioners, this means a continuous learning curve. Lifelong learning will extend to understanding how to best integrate AI tools into their daily workflow to enhance diagnostic accuracy, treatment planning, and patient care. The goal isn’t to replace the doctor, but to empower them with capabilities previously unimaginable, transforming the role of the physician from sole diagnostic arbiter to a highly skilled orchestrator of human and artificial intelligence, all for the benefit of the patient. (See: CDC report on diagnostic accuracy.)

The curriculum transformation in medical schools will be profound. We’ll see courses on “Clinical AI Integration,” “Data Interpretation for AI-Assisted Diagnosis,” and “Ethical AI in Healthcare.” The emphasis will shift from pure knowledge recall, which AI can now largely handle, to critical thinking, problem-solving in ambiguous situations, and mastering the art of patient communication and empathy. Medical students will learn to view AI as an extension of their cognitive abilities, a tool to augment their diagnostic reasoning rather than replace it. For seasoned doctors, professional development programs will become essential, offering practical training on how to seamlessly incorporate AI in disease diagnosis into their existing practices, interpret AI-generated insights, and communicate AI-informed diagnoses to patients in an understandable and reassuring way. This evolution ensures that the human element of medicine remains central, even as technology fundamentally reshapes its capabilities.

10. The Landscape of AI in Disease Diagnosis: Beyond MAI-DxO

While Microsoft’s MAI-DxO is making headlines, it’s important to recognize that the field of AI in disease diagnosis is a vibrant and competitive one, with numerous other players contributing significant innovations. MAI-DxO is a prime example of a large language model (LLM) applied to diagnostics, but other AI approaches are also showing immense promise.

For instance, deep learning models, particularly convolutional neural networks (CNNs), have revolutionized image-based diagnostics. Systems like Google’s DeepMind have demonstrated expert-level accuracy in detecting diabetic retinopathy from retinal scans, often outperforming ophthalmologists. Similar AI applications are being developed for mammography to detect breast cancer, for identifying skin cancers from dermatoscopic images, and for analyzing pathology slides to detect various malignancies. These systems excel at pattern recognition within complex visual data, a task where human eyes can sometimes miss subtle indicators due to fatigue or the sheer volume of images. For more context, see The Hepatitis B Cure.

Another area of intense development is the use of AI for genomics and personalized medicine. By analyzing an individual’s genetic makeup, AI can predict disease susceptibility, identify optimal drug therapies, and even suggest preventative measures tailored to a person’s unique biology. IBM Watson Health, despite some past challenges, has explored AI’s role in oncology, assisting doctors in sifting through vast amounts of research to recommend personalized cancer treatments based on a patient’s genetic profile and tumor characteristics.

Furthermore, AI is being applied to wearable sensors and continuous monitoring devices. These systems can track vital signs, activity levels, and other biometric data, using AI to detect early warning signs of conditions like heart failure, sepsis, or even impending diabetic crises, often before symptoms become apparent to the patient or a clinician. This proactive, preventative approach powered by AI in disease diagnosis has the potential to shift healthcare from reactive treatment to predictive intervention.

The competition and collaboration among these diverse AI applications are pushing the boundaries of what’s possible, creating a rich ecosystem where different AI methodologies address various diagnostic challenges. MAI-DxO’s strength lies in its comprehensive reasoning across multiple data types, but it’s part of a much larger, global effort to harness artificial intelligence for a healthier future.

11. Real-World Examples and Case Studies: AI in Action

It’s one thing to discuss theoretical capabilities and lab results; it’s another to see AI in disease diagnosis making a tangible difference in real-world scenarios. While MAI-DxO is still in clinical trials, numerous other AI systems are already integrated into clinical practice globally.

One compelling example comes from the field of cardiology. AI algorithms are now routinely used to analyze electrocardiograms (ECGs) with incredible speed and accuracy. These systems can detect subtle abnormalities that might be missed by the human eye, predicting conditions like atrial fibrillation or even the risk of sudden cardiac arrest. In some emergency departments, AI-powered ECG analysis provides immediate feedback, helping clinicians make faster, more informed decisions during critical moments. This has led to quicker interventions and improved patient outcomes for heart-related conditions.

In ophthalmology, AI tools are transforming screening programs for diabetic retinopathy. In countries with high diabetes prevalence, there aren’t enough ophthalmologists to screen every patient regularly. AI systems can analyze thousands of retinal images quickly, flagging those with potential disease for human review, thus drastically increasing screening efficiency and preventing vision loss for countless individuals. This is particularly impactful in remote or underserved areas where access to specialists is limited.

Radiology has also seen significant integration. AI models assist radiologists in identifying suspicious lesions on X-rays, CT scans, and MRIs. For example, AI can highlight tiny lung nodules that might indicate early-stage cancer, or pinpoint subtle fractures that could be overlooked in a busy workload. While the final diagnosis still rests with the radiologist, the AI acts as an invaluable second pair of eyes, reducing diagnostic errors and improving the speed of interpretation.

Even in mental health, AI is showing promise. Natural Language Processing (NLP) models can analyze patient narratives, therapy session transcripts (with proper consent and anonymization), or even speech patterns to identify indicators of depression, anxiety, or other mental health conditions, providing early alerts and helping clinicians tailor interventions. These examples underscore that AI in disease diagnosis isn’t a distant dream; it’s a rapidly evolving reality already impacting patient care across various medical specialties.

12. The Future Trajectory: What’s Next for AI in Diagnostics?

Looking ahead, the evolution of AI in disease diagnosis promises even more transformative changes. We’re likely to see several key trends emerge and accelerate in the coming years.

First, **multi-modal AI integration** will become standard. Current AI systems often specialize in one type of data – images, text, or genomic sequences. Future systems will seamlessly integrate all these data sources, creating a holistic patient view. Imagine an AI that combines your genetic predispositions, real-time wearable data, electronic health records, imaging results, and even environmental factors to provide an incredibly precise and dynamic risk assessment and diagnostic pathway. This comprehensive approach will lead to even more accurate and personalized diagnoses. For more context, see Groundbreaking Hibsago Hepatitis B Treatment. (See: ScienceDirect article on AI in diagnostics.)

Second, **predictive and preventative diagnostics** will gain prominence. Instead of diagnosing a disease once symptoms appear, AI will increasingly predict the likelihood of developing certain conditions years in advance. This could involve analyzing subtle shifts in biomarkers, lifestyle data, and genetic markers, allowing for early interventions that prevent disease onset altogether. This shift from “sick care” to “well care” could revolutionize public health.

Third, **explainable AI (XAI)** will be crucial for building trust. As AI systems become more complex, understanding *why* they make a particular diagnosis is essential for doctors to trust and validate their recommendations. Future AI in disease diagnosis tools will need to provide clear, human-understandable explanations for their reasoning, citing the evidence and data points that led to their conclusions. This transparency will be vital for widespread adoption and for addressing ethical concerns about “black box” algorithms.

Finally, **democratization of advanced diagnostics** will be a major goal. The aim isn’t just to make healthcare better for those in well-funded urban hospitals, but to extend high-quality diagnostics to remote areas and developing countries. Portable, AI-powered diagnostic devices, perhaps integrated with smartphones, could bring expert-level diagnostic capabilities to places where medical specialists are scarce, significantly reducing global health disparities. The trajectory of AI in disease diagnosis points toward a future where diagnostics are more accurate, personalized, predictive, and globally accessible.

Frequently Asked Questions (FAQ) about AI in Disease Diagnosis

Q1: Will AI replace doctors entirely?

No, it’s highly unlikely AI will replace doctors entirely. While AI systems like MAI-DxO show superior diagnostic accuracy in specific test scenarios, medicine involves much more than just diagnosis. Human doctors provide empathy, communication, complex decision-making in ambiguous situations, surgical skills, and hands-on patient care that AI cannot replicate. AI is expected to act as a powerful assistant, augmenting doctors’ capabilities and allowing them to focus more on the human aspects of care.

Q2: How does AI achieve such high diagnostic accuracy compared to humans?

AI achieves high accuracy by leveraging several advantages. It can process and analyze vast amounts of data (millions of medical records, research papers, images) at speeds impossible for humans. AI doesn’t experience fatigue or cognitive biases, and its pattern recognition capabilities, especially in deep learning models, are exceptional. For LLM-based systems like MAI-DxO, the ability to understand complex language, reason across diverse medical information, and synthesize data from various sources contributes to its superior diagnostic performance.

Q3: What types of diseases is AI best at diagnosing?

AI currently excels in areas with large datasets and clear patterns. This includes image-based diagnostics for conditions like certain cancers (e.g., breast, lung, skin), diabetic retinopathy, and various neurological disorders from MRI scans. AI is also very effective in analyzing genomic data for personalized medicine and identifying rare diseases. As AI evolves, its capabilities are expanding to complex, multi-symptom conditions, as demonstrated by MAI-DxO.

Q4: What are the main ethical concerns with using AI for diagnosis?

Key ethical concerns include accountability for diagnostic errors made by AI, ensuring equitable access to prevent a two-tiered healthcare system, maintaining data privacy and security of highly sensitive patient information, and addressing potential algorithmic bias. Bias could lead to less accurate diagnoses for certain demographic groups if the AI’s training data was not representative. Transparency and explainability of AI’s reasoning are also crucial for trust.

Q5: How will AI integration impact medical education?

Medical education will need to adapt significantly. Future doctors will require “AI literacy,” understanding how AI systems work, how to effectively use them as diagnostic tools, and how to interpret their outputs critically. Curricula will likely include modules on data science, ethical AI in healthcare, and focusing on human-centric skills like empathy, communication, and complex problem-solving that AI cannot replicate. Lifelong learning for current practitioners will also become more important to stay updated with AI advancements.

Q6: Is AI in disease diagnosis already being used in hospitals today?

Yes, AI is already in use in various capacities within hospitals and clinics. AI-powered systems assist in analyzing medical images (X-rays, CTs, MRIs) in radiology, interpreting ECGs in cardiology, and screening for conditions like diabetic retinopathy in ophthalmology. While a comprehensive diagnostic orchestrator like MAI-DxO is still in clinical trials, specialized AI tools are increasingly integrated into specific medical workflows to enhance efficiency and accuracy.

Q7: How will AI reduce healthcare costs?

AI can reduce healthcare costs primarily by improving diagnostic accuracy and efficiency. By pinpointing diagnoses more quickly and accurately, AI can reduce the need for multiple, often redundant or unnecessary, tests and specialist consultations. This streamlines the diagnostic pathway, reduces trial-and-error, and ensures more targeted use of resources. Early and accurate diagnosis can also