Challenges and applications

Measuring concentrations of steroid hormones – specifically androgens and estrogens – is challenging because they are present at very low concentrations and are structurally very similar. CLI chatted to James Hawley to find out more about these difficulties and how the assays are beginning to be used to measure testosterone levels in prostate cancer patients undergoing androgen deprivation therapy.

We all know that testosterone and estrogen are steroid hormones – can you give us a brief overview of the importance of them, please?

Testosterone and estradiol (estrogens) are very closely linked. They are produced in a pathway that we refer to as steroidogenesis, which is mainly localized in the adrenal glands. Around puberty these hormones will also start to be produced from the sexual glands in females and males, the ovaries and testes, respectively. So at puberty we see significant increases in the concentrations of these hormones which help to develop secondary sexual characteristics: in males testosterone helps stimulate facial hair growth, libido, increased muscle mass and bone strengthening; in females, estradiol helps to initiate and maintain the menstrual cycle premenopausal females, as well promoting bone and muscle health.

These hormones also have important roles in several conditions related to dysregulated hormones.

In females, we may see polycystic ovary syndrome (PCOS) when androgen production is in excess. In males we may be more concerned with hypogonadism, where patients aren’t producing enough androgen/testosterone. Similarly, these hormones are growth factors, so they promote cell growth, this can be proble-matic with certain types of tumours, (i.e. prostate cancer). Also, in females, breast cancer cells can be receptive to estrogen, so if the levels are not suppressed, these can contribute to tumour growth. Therefore, it is important that the concentration of these hormones are monitored in the appropriate clinical context.

It’s crucial to remember that women need testosterone and men need estradiol. Both are important and closely related, as estradiol is produced from testosterone by the action of aromatase. This can become important in certain metabolic conditions like obesity, as fat cells express aromatase, so, if you have excess fat cells then you can produce more estradiol.

Prostate cancer is driven by testosterone Androgen deprivation therapy reduces the level of testosterone to minimize tumour growth and monitoring testosterone levels is required to demonstrate treatment efficacy (Adobe Stock)

Why is measuring the levels of these hormones important?

There are several critical questions that we may look at. In pediatric populations, we may be looking for precocious puberty. This is if sexual development has started prior to the age of eight and we’re seeing increased levels of androgens or estrogens that may be driving sexual development. Something that could be concerning even prior to puberty is congenital adrenal hyperplasia, where you can see elevated androgens which can cause the formation of ambiguous genitalia. This can be a serious condition in the first few days of life. Moving into adulthood, we may investigate females for conditions like PCOS, which can be a difficult condition to diagnose. In males, primarily we see hypogonadism, where insufficient concentrations of androgens are produced and patients may see low libido or erectile dysfunction. There are also symptoms linked to hypogonadism that may prompt measuring these hormones, for example, stunted growth, poor development of secondary sexual characteristics, low muscle mass, these may indicate testosterone deficiency. In addition, conditions like gynecomastia (the development of breast tissue in males) might indicate the measurement of estradiol and testosterone to help identify the underlying cause.

Then we fall into the realms of therapeutic drug monitoring for conditions like breast cancer where patients may be given particular medications to suppress the production of estradiol, which, as it is also a growth factor, can promote tumour growth in hormone-receptor positive cases. There are certain medications that are prescribed to reduce estradiol concentrations and so help to limit tumour growth. Similarly with testosterone and prostate cancer therapy, androgen deprivation therapy (ADT) can be administered to limit the production of testosterone. Again, the objective is to suppress testosterone production and so reduce the exposure of the growth factor which is driving the tumor production. Here, we need to measure testosterone concentrations to ensure that the ADT is working and the dose is correct.

We also increasingly see females that are peri- or post-menopausal, wishing to supplement with testosterone to maintain their bone health and libido. It’s important to monitor this as testosterone replacement may make them prone to polycythemia (increased red cell production), which can lead to blood hyperviscosity. There is a balance to be achieved with replacing and over replacing testosterone in females.

How is assaying these hormones usually done and what are the challenges?

Undoubtedly, the most common platform for measuring these hormones and other steroid hormones is immunoassay.

Specificity
Steroid hormones are small. Estradiol is 272 Da and testosterone is 288 Da, which means that they are not conducive to multi-site immunoassays. Normally, with larger proteins, you have the option of multiple targets which are antibodies can bind to allowing for a sandwich ELISA. Generally, this helps as the more antibodies that can target the analyte, the more specific the assay become, and it may also produce a better signal thereby improving sensitivity. However, because testosterone and estradiol are so small, they’re only really amenable to competitive, single-site antibody immuno-assays. This is problematic for steroid hormones because they are all very similar in structure. Steroid hormones – progesterone, estradiol, cortisol, cortisone, testosterone, androstenedione, dihydrotestosterone, etc – are all closely related. They are all synthesized from cholesterol and have the same basic steroid framework, consisting of three six-membered rings (A, B, C) and one five-membered ring (D). Therefore, achieving good specificity in an immunoassay is very difficult.

Sensitivity

Achieving sensitivity in immunoassays is also challenging because these molecules are not present in high concentrations. For testosterone, we’re working with nanomolar concentrations and as the concen-trations differ from females to males, we typically need a working range of 0.1 nmol/L up to around 50 nmol/L, which is quite wide. However, estradiol is present at the picomolar range, and certainly for hormone-receptor positive breast cancer patients, the treatment target is to reduce estradiol concentration to less than 3 pmol/L per litre, according to the guidelines, and that is really challenging to measure. Most immunoassays for estradiol can only really get to 80–92 pmol/L as a lower limit of quantitation, so these assays are not capable of measuring those low levels.

Cross reactivity

As we have mentioned, specificity is an issue, because the structure of the steroid hormones are so similar, antibodies have a tendency to cross-react with structural homologs. Testosterone, its precursors androstenedione and dehydroepiandrosterone (DHEA) are structurally similar; for the estrogens, estradiol and estrone are also very similar and cross reactivity can occur if the assay is not well designed. Additionally, exogenous hormones may also cause assay interference.

However, it is extremely challenging for immunoassay manufacturers to produce an assay that is accurate and specific. Over the years, a lot of work has been done to show the discordance between immuno-assays from different providers.

Mass spectrometry

These challenges can be circumvented by using mass spectrometry (MS). We use liquid chromatography (LC) in line with tandem MS (LC-MS/MS). The LC is used to separate isobars (molecules of the same mass, such as testosterone and epitestosterone), which helps prevent isobaric interference. MS alone would not be able to distinguish these two molecules apart, but because their functional groups have different orientations, we can fine tune the chromatography to ensure that we separate these isobars, which adds an invaluable layer of specificity. Similarly, with estradiol, there is 17-alpha estradiol and 17-beta estradiol, which differ just by the stereochemical orientation of the hydroxyl group on carbon 17. It is really difficult for immunoassays to distinguish between these isomers but we can with LC-MS/MS.

After the chromatographic separation, the molecule enters the mass spectrometer and we can set the parameters to detect a select mass. For measuring testosterone, we can say we want a parent of 288 Da, so it goes through a quadrupole set to filter only masses of 288 Da, which removes all the other steroids that are not 288 Da. The testosterone molecules then enter a collision cell where they are bombarded with an inert gas, this causes the steroid to break apart into fragments of 97 Da or 109 Da. The next quadrupole is set to scan for 97-Da fragments (we can also scan for 109-sized fragments to look at the ratio between the two). So, with this approach, we have achieved chromatographic separation, set the mass spectrometer to scan for a specific transition of testosterone (288 > 97 Da. We then generate a result which we can be confident is accurate.

Challenges of LC-MS/MS

Standards and calibration

There is a misplaced belief that if you measure something using LC-MS/MS, you always getting the right answer; however, you can have badly designed MS assays which don’t produce good results.
If the chromatography is not quite right, you may suffer from co-eluting interferences and/or matrix effects. The choice of internal standard is also critical, ideally, you need something that co-elutes with the analyte of interest, carbon-13 internal standards are routinely available now for testosterone and estradiol and should be used in preference to deuterated internal standards or analogues. Importantly, good calibrators and the use of independent quality controls are essential to ensure the assay is accurate and precise; If you are making in-house standards from powdered or liquid material, you need to know the provenance of that material, where possible, certified reference material is typically superior as you can show traceability to SI units. Commercial calibrators are also increasingly available for MS assay and represent a viable option.

Assay design and limit of quantitation

Additionally, you need to consider the design of the assay and the limit of quantitation. This becomes especially important when measuring estradiol concentrations. There are not many labs that can measure as little as 3 pmol/L, and for achieving this quite a large sample may be required.

Specialized MS instrumentation

The detection of these very low concentrations of estradiol is done on a very high-end MS instrumentation. Most steroids ionize very well in what we call positive ionization mode, meaning that an M+1 ion of the molecule is formed by the addition of a charged hydrogen ion. Thus, the mass of testosterone, which is 288 Da, becomes 289 Da because a charge is needed for the molecules to pass through the mass spectrometer.

Estradiol, on the other hand, is only really amenable to MS in negative ionization mode. You don’t always get as good a signal in negative ionization mode as in positive, but some manufacturers do produce instruments that are configured really well for negative ionization mode and that helps with some of the challenging analytes that need negative ionization mode, such as estradiol and aldosterone.

Derivatization

It is possible to circumvent the need for negative ionization mode in MS by creating a derivative (derivatizing) estradiol, by exposing it to conditions where it will form an adduct, and then using the mass spectrometer to scan for estradiol and that adduct. This can increase sensitivity by virtue of making it amenable to positive ionization mode. However, there is a school of thought that derivatization is not always the best practice with some steroids because usually when you’re creating a derivative, the sample is subjected to quite severe conditions (e.g. heat) and you don’t always know what impact that is having. For example, because the structures of the steroids are so similar, it may be that some steroids alter their conformation during derivatization to form estradiol, so we may be causing effects that we don’t know about. If possible, it’s generally best to avoid derivatization and to keep the number of steps in the process to a minimum to limit any uncertainty around the final result.

What are the potential improvements that could be made?

Clinical awareness and funding
I think increased clinical awareness would signify the biggest shift. We need the end users –the clinicians – to be aware of how that result is produced and cognizant of the assay that is being used.
Not all hospital laboratories have direct access to LC-MS/MS, and it can be expensive, so it’s not always ideal to send specimens away to a specialist centre. The requesting clinician, therefore, needs some awareness of what is appropriate and what is not appropriate. If a patient is taking fulvestrant [an estrogen receptor antagonist used for treating estrogen-receptor positive breast cancer] and you want to measure estradiol, you need to be aware that this could be a problem on some platforms and that if you want an accurate result it needs to be referred for a MS assay. So awareness is key.

Automated MS
Potentially one development that will come into play in the next 5 to 10 years is automated MS analysis. Some manufacturers are now looking at integrating MS technology into their automated tracked systems, This will absolutely increase the availability of MS thereby helping to provide the most clinically appropriate assay for that patient’s sample.

Reducing the lower limit of quantitation
Assaying for estradiol is always going to be challenging, but there are still improvements that can be made for measuring testosterone. As we have mentioned, with therapeutic drug monitoring for prostate cancer patients on ADT it would be great to be able to reduce the lower limit of quantitation even further and to really go into the picomol per litre range and see if the ADT
is fully suppressing testosterone levels, just as we are now for estradiol in hormone-receptor positive (HR+) breast cancer.

Improving therapeutic drug monitoring of prostate cancer patients on ADT
Not much work has been done looking at testosterone levels in males taking ADT. We know that immunoassays don’t work well in that low nanomolar range, so we need to increase awareness that if patients on ADT still have elevated testosterone concentrations, it may be by virtue of cross reactivity in the immuno-assay and a MS assay would be more appropriate.

The structures of the estrogens (a) and androgens (b) are all based on the A, B, C, D ring backbone and are very similar (Adobe Stock)

Examining the impact of 11-oxygenated androgens

One of the exciting areas of steroid hormone research has been the emergence of 11-oxygenated androgens, such as 11 ketotestosterone, which are produced from a slightly different pathway to classic steroidogenesis. 11-ketotestosterone has the same affinity for the androgen receptor as testosterone and could potentially drive some forms of prostate cancer. It may be that the non-specificity of the immunoassays is providing information about these other forms of testosterone that we are unintentionally assaying as a result of the antibody cross reactions. In order to properly understand the im-
pact of the ketotestosterone on prostate cancer we need to measure it and some of the other 11-oxygenated androgens by MS to see if these are cross-reacting or not. In the meantime, we know that testosterone is a major androgen involved in prostate cancer and we should focus on trying to measure that as accurately as possible, before perhaps expanding the MS or immunoassays to detect ketotestosterone.

Analysis of free hormone and use of saliva as a specimen
It’s important to appreciate how these hormones circulate as well. Really, what we want to monitor is the free hormone. However, testosterone is avidly bound to its carrier protein sex hormone-binding globulin (SHBG) and also to a lesser extent albumin, so only around 2–3% of testosterone is actually free and this is what is considered

physiologically active. In the laboratory we measure total testosterone, so we don’t really have an appreciation for the free form. Similarly with estradiol, this isn’t bound as avidly to SHBG protein as testosterone, but it’s still significantly bound to SHBG and albumin. Ideally, we would monitor the free form as this is physiologically active, but this is technically demanding with the assays that we use, so it’s not always possible. Equations have been developed for testosterone to calculate the free quota, which take into account the binding capacity of SHBG and albumin for testosterone. However, it may be better to start to move away from considering total testosterone and focus more on the free (biologically active) quota. A lot of pioneering work has been done in saliva and there are now salivary assays for testosterone available, which measure the free quota. However, more work still needs to be done in this area.

It’s more difficult for estradiol, of course, because of the lower order of magnitude that it circulates in. Testing for picomolar quantities in saliva is a lot more challenging. However, there is a big push with the saliva hormone testing for investigating Cushing syndrome or adrenal insufficiency. This obviously can make it easier for patients – not many people love having their blood taken and saliva testing can usually fit into people’s lifestyles better – but, again, more work needs to be undertaken to solidify the evidence base. It would be interesting to investigate and monitor free testosterone in prostate cancer patients and free estradiol in breast cancer patients.

Multiplexing
Additionally, we could also look at expanding steroid panels. One of the advantages of MS over other techniques is that you can quite readily integrate additional analytes into assays – so called ‘multiplexing’ – thereby measuring multiple hormones with just one sample. The aim would be to quantitate the classic androgens (testosterone and androstenedione) as well as the 11-oxygenated androgens (such as 11-hydroxyandrostenedione and 11-ketotestosterone) all from that one serum sample.

AI and risk stratification
Then there’s always the potential of incorporating some form of machine learning into the analysis of the data to try and stratify whether that patient’s at risk or whether they may need treatment titrating, which moves towards the idea of precision medicine.

It’s an exciting time for this challenging field!

The interviewee

James Hawley MSc, FRCPath
Principle Clinical Scientist

Wythenshawe Hospital, Manchester University NHS Foundation Trust, Manchester, UK and Laboratory of Medical Science, Medical Research Council, London, UK

Email address: james.hawley@mft.nhs.uk

Bibliography
1. Mauvais-Jarvis F, Bhasin S. Metabolic Messengers: testosterone. Nat Metab. 2026;doi:10.1038/s42255-025-01431-6 (https://doi.org/10.1038/s42255-025-01431-6).
2. Hevener AL, Correa SM. Metabolic Messengers: oestradiol. Nat Metab. 2025;7(6):1114–1122 (https://doi.org/10.1038/s42255-025-01317-7).
3. Elhassan YS, Hawley JM, Cussen L, Abbara A, Clarke SA et al. Society for Endocrinology Clinical Practice Guideline for the Evaluation of Androgen Excess in Women. Clin Endocrinol (Oxf). 2025;103(4):540–566 (https://doi.org/10.1111/cen.15265).
4. Samuel E, Chiang C, Jennens R, Faulkner D, Francis PA. Fulvestrant falsely elevates oestradiol levels in immunoassays in postmenopausal women with breast cancer. Eur J Cancer. 2020;126:104–105 (https://doi.org/10.1016/j.ejca.2019.10.015).
5. Owen LJ, Monaghan PJ, Armstrong A, Keevil BG, Higham C et al. Oestradiol measurement during fulvestrant treatment for breast cancer. Br J Cancer. 2019;120(4):404–406
(https://doi.org/10.1038/s41416-019-0378-9).
6. Cuzick J, Chu K, Keevil B, Brentnall AR, Howell A et al. Effect of baseline oestradiol serum concentration on the efficacy of anastrozole for preventing breast cancer in postmenopausal
women at high risk: a case-control study of the IBIS-II prevention trial. Lancet Oncol. 2024;25(1):108–116 ( https://doi.org/10.1016/s1470-2045(23)00578-8). ).
7. van Rossum HH, van Winden LJ, Bergman AM, van der Poel HG. Current routine testosterone immunoassays are unsuitable for lowering the general castration cutoff recommendation to <0.7 nmol/l (20 ng/dl). Eur Urol Open Sci. 2023;48:98–100 (https://doi.org/10.1016/j.euros.2022.12.009).
8. Fanelli F, Peitzsch M, Bruce S, Cantù M, Temchenko A et al. Report from the HarmoSter study: different LC-MS/MS androstenedione, DHEAS and testosterone methods compare well; however, unifying calibration is a double-edged sword. Clin Chem Lab Med. 2024;62(6):1080–1091 (https://doi.org/10.1515/cclm-2023-1138).