What Is Ground Glass Opacity and When Is It Serious?

Ground-glass opacity is a term radiologists use to describe a hazy, translucent area on a CT scan of the lungs, where the underlying lung structures (blood vessels, airways) are still visible through the cloudiness rather than completely obscured. It is not a diagnosis in itself but a visual pattern that can reflect dozens of different conditions, from a passing infection to early-stage lung cancer. Whether a particular ground-glass opacity deserves concern depends heavily on context: what else is happening clinically, whether the opacity sticks around or vanishes on repeat imaging, and what it looks like in detail.

What You Are Actually Seeing on the Scan

Normal lung tissue on a CT scan appears almost black because the air-filled spaces absorb very little of the X-ray beam. When something partially fills or thickens those air spaces or the tissue between them, the area becomes slightly brighter but not fully white. That in-between haziness is ground-glass opacity, named because it resembles frosted or ground glass. A fully white area, by contrast, is called consolidation and means the air spaces are completely filled with fluid, cells, or other material.

The causes behind this partial filling are wide-ranging. Fluid from pulmonary edema, inflammatory cells from pneumonia, abnormal proteins from certain immune reactions, or early tumor cells can all produce the same hazy look. As one review put it, the underlying causes include edema, airspace and interstitial pneumonia from various organisms, non-infectious inflammatory conditions, and tumor growth.¹ That is what makes GGO a challenge: the pattern alone does not tell you the cause.

Transient Causes That Usually Are Not Serious

A large share of ground-glass opacities that show up on CT scans turn out to be temporary. Infections are the most common culprit. Viral pneumonias, bacterial infections, and fungal lung disease can all produce GGO. COVID-19 brought this pattern into public awareness because bilateral ground-glass opacities were among the hallmark CT findings in that infection, though GGO is not specific to any one virus.² Other respiratory viruses, Pneumocystis pneumonia in immunocompromised patients, and even some bacterial infections produce similar-looking haziness.

Allergic or eosinophilic lung reactions also commonly show up as ground-glass opacity. One study found that a high blood eosinophil count was strongly associated with GGO lesions that shrank or disappeared on follow-up, suggesting that eosinophilic lung inflammation was likely responsible in those cases.³ Pulmonary edema from heart failure, drug reactions, and bleeding into the lung can produce GGO as well. In all these situations, the opacity typically resolves once the underlying problem is treated.

This is why a repeat CT scan after a few weeks or months is often the most important first step when GGO is found incidentally. If the hazy area disappears, the clinical concern drops substantially. In that same study, roughly 38% of pure ground-glass lesions and nearly 49% of mixed lesions shrank or vanished on follow-up imaging.³

When GGO Persists and What That Means

The real clinical concern begins when a ground-glass opacity does not go away. A GGO that remains stable across two or more CT scans separated by months is called persistent, and persistent GGO carries a meaningfully higher chance of being an early malignancy. The persistence itself is a red flag: infections and inflammatory causes tend to clear, while precancerous or cancerous lesions do not.⁴

Radiologists and pulmonologists pay close attention to certain features when evaluating a persistent GGO nodule. Size matters, but so does whether the opacity is “pure” (entirely hazy, with no solid center) or “part-solid” (a mix of haze and a denser core). Part-solid nodules carry a higher probability of being malignant compared to pure ground-glass nodules. In the study mentioned above, about 19% of pure GGO lesions that persisted turned out to be precancerous or cancerous, while roughly 30% of mixed-density GGO lesions were malignant.³ That gap matters for how aggressively a doctor will pursue further workup.

Another warning sign is growth, whether the GGO gets larger or develops a new solid component within it. An opacity that was once purely hazy and later develops a denser core in the center is behaving more suspiciously. The more extensive the solid portion, the higher the probability of malignancy and the worse the outlook if it does turn out to be cancer.⁴

The Connection to Lung Cancer

When persistent GGO nodules are cancerous, they are overwhelmingly a specific type: lung adenocarcinoma or its precursor lesions. These range from atypical adenomatous hyperplasia (a precancerous stage) through adenocarcinoma in situ and minimally invasive adenocarcinoma to fully invasive adenocarcinoma. GGO-predominant lung cancers represent the early, slow end of the lung cancer spectrum.⁵ This matters because these tumors behave very differently from the aggressive lung cancers most people picture.

Screening with low-dose CT has made GGO findings far more common, and the reported cancer rate among persistent GGO nodules has been as high as 63% in some series.⁶ That number sounds alarming, but it needs context. Many of these cancers are at the earliest possible stage and grow extraordinarily slowly. Pure ground-glass nodules that do grow have median volume-doubling times measured in years, not months. One study found a median doubling time of about 769 days for growing pure GGO nodules.⁷ Another study using detailed volumetric tracking reported even longer times, with a median of roughly four years.⁸ For perspective, many aggressive solid lung cancers double in volume in weeks or a few months.

This glacial growth rate is why the management approach for pure GGO nodules is usually surveillance rather than immediate surgery. The Fleischner Society, which publishes widely used guidelines for managing incidentally found lung nodules, recommends specific follow-up intervals depending on size and characteristics of the nodule.⁹ For very small pure GGO nodules, follow-up imaging may not even be recommended. For larger or part-solid nodules, closer surveillance and possible biopsy are standard.

How Doctors Decide What to Do Next

If a GGO nodule persists and raises enough concern, several tools help determine whether it is malignant. CT-guided needle biopsy is one option and performs reasonably well, with one study reporting overall diagnostic accuracy of about 91%.¹⁰ The tradeoff is a notable risk of pneumothorax (about 18% in that study, though most cases were minor and resolved without treatment).

PET scans, which detect metabolically active tissue, are commonly used for evaluating suspicious lung nodules but have a real limitation with ground-glass opacities. The metabolic activity picked up on PET tends to correlate with lesion size and with how much solid tissue is present: the more ground glass and less solid component, the less metabolic signal.¹¹ In purely ground-glass nodules, PET scans often cannot distinguish cancer from inflammation at all, because the metabolic activity of malignant and benign pure GGO nodules is essentially the same.¹² This is a common source of confusion for patients who assume a “negative PET” means a GGO nodule is safe. It may simply mean the lesion is too small or too airy for PET to detect.

In practice, the most reliable diagnostic approach for persistent GGO nodules remains serial CT imaging over time. Growth or the development of a solid component triggers either biopsy or surgical resection. Many specialists will watch a stable pure GGO nodule for years before recommending intervention, because the biological behavior of these lesions often allows that window.

Outcomes After Surgery

When a GGO-predominant lung cancer does need surgery, the outcomes are strikingly good compared to other forms of lung cancer. Surgeons increasingly use sublobar resection, a smaller operation that removes less lung tissue than the traditional lobectomy, for these slow-growing tumors. A prospective study that followed patients for a median of ten years after sublobar resection for GGO-dominant lung cancer found zero recurrences during the study period and 100% disease-specific survival at ten years.¹³

A larger study of over 300 patients who had sublobar resections for GGO-dominant tumors confirmed similarly excellent results: ten-year relapse-free survival was about 99%, and ten-year overall survival was about 99% as well, with only a single local recurrence at the resection margin across the entire cohort.¹⁴ For patients with small, peripheral, GGO-dominant tumors, sublobar resection with adequate margins appears to offer excellent long-term control.¹⁵

These numbers are dramatically better than lung cancer outcomes overall. They reflect the indolent biology of GGO-predominant adenocarcinomas and are part of the reason many experts advocate for careful surveillance rather than rushing to the operating room for every persistent GGO.

The Overdiagnosis Problem

The excellent surgical outcomes carry an uncomfortable flip side: some of these tumors may never have caused harm even without treatment. This is the overdiagnosis dilemma. As low-dose CT screening has expanded, particularly in East Asian populations where lung cancer in never-smokers is more common, the number of detected GGO nodules has surged. Many of these, especially pure GGOs, may never progress to invasive cancer during a person’s lifetime.¹⁶

Overdiagnosis leads to overtreatment: surgery with its attendant risks, loss of lung tissue, psychological burden, and cost. An expert consensus statement acknowledged the core tension: clinicians face a dilemma between timely intervention and avoiding unnecessary treatment, while patients often experience follow-up anxiety and confusion about what to do.¹⁷

The anxiety itself is a real clinical problem. Among patients with GGO findings awaiting surgery, about a quarter reported significant preoperative anxiety and roughly 18% met criteria for depression. Having multiple GGO nodules was an independent risk factor for anxiety, and the psychological distress was associated with worse pain scores after surgery and lower quality of life.¹⁸ If you have been told you have a GGO on a scan and find yourself consumed by worry, you are not alone and that reaction is well documented.

Genetic Clues Inside the Nodule

Researchers have found that the molecular makeup of GGO-associated lung adenocarcinomas differs in interesting ways from solid lung tumors. Mutations in the EGFR gene are particularly common. One study found that about 48% of resected GGO lesions harbored EGFR mutations.¹⁹ The frequency of EGFR mutation varied with how advanced the tumor was: adenocarcinoma in situ (the earliest cancerous stage) had the lowest rate at about 31%, while more advanced subtypes had much higher rates.²⁰

That same study on EGFR found that the mutation appeared to accelerate how quickly a GGO progressed on imaging, meaning EGFR-positive GGO nodules tended to grow or develop solid components faster than EGFR-negative ones.¹⁹ For now, genetic testing is generally done after a nodule has been surgically removed, not as a routine part of surveillance. But these molecular patterns are starting to inform how aggressively doctors follow certain patients.

Why Measuring Tiny GGO Nodules Is Harder Than It Sounds

One underappreciated challenge with GGO surveillance is measurement accuracy. Because ground-glass nodules have indistinct, fading borders rather than the sharp edges of a solid lump, measuring their exact size is trickier than it looks. Computer-aided volume measurements work well for nodules about 5 mm or larger, with errors in the single-digit percentages.²¹ But for very small nodules around 3 mm, the measurement error can be enormous, ranging from 51% to 85% in one study.²¹

The good news is that the CT scanning technique itself does not seem to distort GGO measurements much. A phantom study found that low-dose scanning protocols and different image-processing methods produced volume measurements with consistently low error, averaging less than 6%.²² The limitation is really about the tiny size of some nodules and the inherent blurriness of ground glass, not the scanner settings. This is one reason guidelines tend to recommend longer follow-up intervals for very small GGO nodules: a reported “growth” of 1 mm in a 3-mm nodule could easily be measurement noise.

Artificial Intelligence in GGO Assessment

Radiologists are increasingly testing AI tools to help predict whether a ground-glass nodule is invasive cancer or something less threatening. The challenge is distinguishing minimally invasive adenocarcinoma from invasive adenocarcinoma on imaging alone, since the two can look very similar on CT but carry different surgical implications. A 2025 study found that combining deep learning with traditional image-feature analysis achieved an accuracy metric of about 0.90 for distinguishing these subtypes, outperforming either method used alone.²³ An earlier study using a similar combined approach reported accuracy above 0.91 for identifying high-grade patterns within adenocarcinomas presenting as GGO.²⁴

These tools are not yet standard in clinical practice, and no AI model has replaced the judgment of an experienced radiologist in this context. But the research is moving fast. If the accuracy holds up in large multi-center trials, AI-assisted reads could eventually help stratify which GGO nodules need closer follow-up and which can safely be watched at longer intervals, reducing both unnecessary biopsies and the anxiety that comes with uncertain surveillance schedules.