Lesion vs. Injury

SNOMED Groups Translation User Group
1

Dear All,

We currently have a lot of cases of duplicate PTs in French for pairs of the type “Lesion of xxx”/“Injury of xxx”, translated in both cases as “lésion”. E.g., “lésion de la clavicule” for both Clavicle injury and Lesion of clavicle. In French, the word “lésion” covers both meanings.

Examples of concepts concerned:

  • Clavicle injury (disorder) Lesion of clavicle (disorder)

  • Injury of cornea (disorder) Lesion of cornea (disorder)

  • Injury of tongue (disorder) Lesion of tongue (disorder)

  • Scalp injury (disorder) Lesion of scalp (disorder)

  • Injury of liver (disorder) Lesion of liver (disorder)

  • Injury of phrenic nerve (disorder) Phrenic nerve lesion (disorder)

  • Among many others…

I see that other languages have struggled with this problem, and I would like to know more about the different approaches taken by other countries.

Lesion of radial nerve (disorder)

Injury of radial nerve (disorder)

Danish

læsioner af nervus radialis (plural form of læsion, if I’m correct)

traumatisk læsion af nervus radialis (even though defined as Traumatic or non-traumatic injury (disorder))

Swedish

lesioner i nervus radialis (also plural?)

skada på nervus radialis (more like damage? cognate with Dutch “schade” or injury?)

Norwegian

lesjon i nervus radialis

skade på nervus radialis

Spanish

lesión de nervio radial

lesión traumática del nervio radial (same problem as with Danish)

Dutch

laesie van nervus radialis

letsel van nervus radialis

French

lésion du nerf radial

lésion du nerf radial

As you can see in Dutch, the solution was to use “laesie” for lesion with (in some cases) synonym “afwijkend weefsel” = abnormal tissue, and “letsel” for injury. This is more in line with the usual English definition of “lesion,” with a much more restricted meaning, referring to damaged tissue, which may or may not be caused by an injury.

E.g. Definition of “lesion” by the National Cancer Institute: An area of abnormal or damaged tissue caused by injury, infection, or disease. A lesion can occur anywhere in or on the body, such as the skin, blood vessels, brain, and other organs. Examples of lesions include wounds, ulcers, abscesses, sores, cysts, and tumors. A lesion may be benign (not cancer) or malignant (cancer). [National Cancer Institute|https://www.cancer.gov/publications/dictionaries/cancer-terms/def/lesion.]

Injury, on the other hand, is defined by SNOMED CT as Disorder resulting from physical damage to the body (synonym of concept Injury). So the nuance between the two would be that “lesion” refers to the presence of abnormal or damaged tissue regardless of its cause, and “injury” refers to the abnormality as a result of physical force.

This would explain why:

Open injury of radial nerve (disorder) is subsumed by both Injury of radial nerve (disorder) and Lesion of radial nerve (disorder), since it’s defined with Associated Morphology: Open wound (=tissue abnormality) Due to: Traumatic event (=physical force).

But Injury of radial nerve at forearm level (disorder) is subsumed only by Injury of radial nerve (disorder) (defined as Is a: Traumatic or non-traumatic injury (disorder) (=physical force)), but not by Lesion of radial nerve (disorder), because it does not denote any specific tissue abnormality.

My questions to fellow NRCs are:

  1. What are your thoughts on the Dutch approach of using a term closer to the English definition that allows to convey the English nuance between “lesion” and “injury”?

  2. Is there a word in your local languages with a meaning closer to the English “lesion” or Dutch “laesie”, i.e. limited to “damaged/abnormal tissue”?

  3. In Norwegian and Swedish: does the word skade/skada denote something effectively different than the word for lesion?

For example, in French this approach would imply that “lesion” concepts would need a new purpose-coined term such as “anomalie tissulaire”, and the “lésion” (with syn. “blessure” and “traumatisme” if traumatic) could both be used in “injury” concepts only, which would be an advantage since these terms are effectively synonyms.

Looking forward to discussing this issue at our next meeting.

Kind regards,

Pedro

*****

Side note on French translations:

The solution currently prescribed by the Common French group is:

  • Lesion = lésion

  • Injury = lésion traumatique ou non traumatique (depending on the defining relationships)

This proposal is problematic because it implies semantically that all lesion concepts are supertypes of injury concepts (since “lesion” is more general than “traumatic/non traumatic lesion”, and therefore for each body part, lesion should subsume injury, which isn’t the case.)

2 Likes
2

A note on the Dutch approach: we only use the word ‘laesie’ as a synonym. The preferred term is ‘afwijkend weefsel’ (abnormal tissue), because this is the closest approximation we could find to the English term ‘lesion’.

We avoided laesie as a preferred term because it does often imply an injury, and yet is often used for things that are not injuries as well. Because the term is used very frequently, it is important to add it for findability. So our compromise is to add ‘laesie’ as a synonym to the concept that comes closest; and use the inelegant but unambiguous ‘afwijkend weefsel’ as preferred term.

In the same spirit, we have added ‘arm’ as synonym to ‘upper extremity’. It’s not exactly the same, but there is no better concept available for a frequently used term. This approach differs from the guidelines SNOMED International uses, because the meanings are not exactly the same. If you go down this route, you should advise software vendors to always display the preferred name when supporting SNOMED search.

1 Like
3

Regarding the Norwegian translation, there is a short answer and a long one.

Short: The established translation in bilingual dictionaries is lesion>lesjon and injury>skade.

Long: The fundamental question is what is the conceptual difference between ‘lesion’ and ‘injury’, i.e. what are the essential charasteristics of these two concepts which deliniated them from eachother. We’ve discussed with clinicians whether the axis traumatic and non-traumatic are key differences, but this doesn’t seem to be the case. Lesions can be traumatic and non-traumatic, the same applies to injuries. The National medical encyclopedia in Norway highlights that lesions are localized and delimited changes in body organs or parts of body organs, but we are unsure whether this is an essential difference from an injury.

A relevant side note is the difference between damage and injury. What is the difference between 243978007 |Liver damage (disorder)| and 39400004 |Injury of liver (disorder)|. We struggle with this because both “damage” and “injury” are normally translated as “skade” in Norwegian.

4

The same problem occurs in swedish as injury (“skada”) is one kind of lesion. In medical dictionaries lesion is explained as could be due to violence or disorder, and is a “medical term for harmful influence on tissue or organs”, but damage (“skada”) is a synonym, but with the added note “lesion is used to refer to any damage to the body, from the fracture of a bone to a cut on the skin”.

In Swedish we can use “lesion” and “skada” to make a difference between the two, but it is an artifical difference, and not useful for the users.

1 Like
5

Answer we received from SNOMED a few years ago:
WRT “Lesion” SNOMED has interpreted it to mean a localized and circumscribed area of structurally altered or diseased tissue. While wounds and injuries are often used synonymously, there are some subtle differences. Wounds generally refer to tissue abnormalities (regardless of cause) that involve surface structures such as skin and subcutaneous tissue, ocular structures, etc. Injuries on the other hand my refer to tissue abnormalities occurring anywhere within or on the body. For example, a traumatic rupture of the spleen would be considered an injury, while a traumatic laceration of the arm could be either a wound or an injury, although we have currently represented them as traumatic injuries. Wounds are also usually the result of some external force, whereas some forms of injury may be non-traumatic (e.g. 391179008 |Non-traumatic pneumopericardium (disorder)|). In some cases, it is just a judgement call based on the common clinical usage, although in SNOMED we have attempted to be consistent in the criteria for modeling something as a wound or an injury. " Wat zijn jullie gedachten hierover?

1 Like
6

Thanks, Pedro. This is interesting. I think it would be an advantage if the text definition of “lesion” would be added to for example 52988006 |Lesion (morphologic abnormality)|.

We’ve consulted a medical expert, and this text definition of “Lesion” coincides with the use of the Norwegian “lesjon”. In other words, this is not a problem when translating into Norwegian. However, the difference between injury and damage in such as 39400004 |Injury of liver (disorder)| and 243978007 |Liver damage (disorder)| is unclear for us and, hence, causes problems when translating. Additionally, the synonym “Hepatic trauma” to 39400004 |Injury of liver (disorder)| might be problematic.

7

Thanks Ole.

I’d like to know how often general concepts such as “Lesion of nose” and “Injury of nose” and “Traumatic injury of nose” are actually used in clinical settings. They seem to me mostly useful when browsing up and down the ontology, but I don’t think they have any real use in clinical settings.

Furthermore, if the more specific descendants (e.g. Burn of nose) are fully defined, these rather vague groupers are not necessary to define intensional refsets such as “descendants of lesion of nose”, since the same can be achieved with more refined expressions such as “disorders with associated morphology descendant of lesion/traumatic injury and site descendant or self of nose”.

My question to NRCs is what is the real use of the problematic concepts? Do you agree that they mostly work as groupers and are therefore not useful in clinical settings? Would inactivation cause other problems?

8

Hi Pedro, I had a brief look in the Dutch hospital thesaurus of diagnoses. I cannot find any concepts there with ‘lesion’, with the exception of ‘spinal cord lesion’. There are a lot of concepts for ‘injury’.

So I can see a requirement for the injury concepts, but not for the lesions. I remember we struggled to find a translation for those as well.

2 Likes
9

A brief historical note may help clarify the meaning of lesion in SNOMED. Much of the early SNOMED CT content released in 2002 incorporated terminology from ICD-9, where lesion functioned as a nonspecific descriptor for any structural abnormality of a body part. ICD-10 continued this usage, applying lesion to denote an abnormal area of tissue identified by any method, without implying a specific morphology or cause. ICD-11 retains the same general meaning and uses lesion to indicate a localized structural abnormality seen on imaging, endoscopy, physical examination, or pathology, again without commitment to an underlying diagnosis.

SNOMED CT inherited this broad, non-etiological notion of lesion, and it continues to represent a generalized morphologic abnormality rather than a causal statement. For this reason, lesion is not equivalent to injury or traumatic injury. Injury inherently implies tissue damage, which may be traumatic (caused by an external force) or non-traumatic (arising spontaneously or through internal pathological processes). Traumatic injury, in contrast, explicitly requires an external physical cause. Thus, the three terms—lesion, injury, and traumatic injury—represent distinct semantic categories.

Clinical usage reflects these distinctions. A radiographic “lung lesion” simply denotes a structural abnormality on imaging, without implying etiology. A tendon rupture may be traumatic if produced by external force or non-traumatic if spontaneous or degenerative. In general, lesion should be understood as a broad descriptor for a structural abnormality of a body structure, independent of cause.

2 Likes
10

Have you checked if ICD-11 in your languages also uses lesion in this way? In Swedish that is not the case. Instead “skada” (damage) is used, or a reference to condition, or pares. Only for some codes, lesion is used in Swedish also (8B81.0). Note also that in many cases, lesion is a synonym only, not the actual title chosen.

11

The French version of ICD-11 was most likely translated by machine, as even the basic body parts lack consistency (e.g., “lower leg” has six different translations, and none are right). Therefore, I do not believe that we can rely on the French version to obtain accurate equivalents of lesion and injury.

1 Like
12

Hi @jcase thank you for these clarifications.

  1. “Thus, the three terms—lesion, injury, and traumatic injury—represent distinct semantic categories.” –> Does this mean that all “injury“ concepts defined as caused by a traumatic event should be changed to include the word “traumatic” in their FSNs?

    Just one example (there are many): 212303005 |Open injury, posterior interosseous nerve (disorder)|, with defining relationship Due to > Traumatic event.

  2. From what you say, I can see how there’s a need in English-speaking clinical settings for “lesion” concepts. However, this is not the case for other languages, definitely not for French and Dutch (Belgium’s languages), where the “lesion” concept is useless when there is an injury counterpart for the same body site, and causes a lot of confusion for practitioners, since both lesion and injury translate as the same words in our languages.
    We could use an artificial translation for lesion as a workaround (sth like ‘tissue abnormality’), but this will not solve the issue.

    In this case, the ideal solution would be for to be able to exclude lesion concepts from our national edition. Is there a possibility SNOMED would allow countries to selectively inactivate certain concepts from the international edition? This could also be used to exclude US/UK concepts related to housing types, occupations, environments, etc. that have no use in other countries.

13

Just a quick remark. You might still need some lesion concepts, I think. But they are generally groupers, so maybe you can manage to avoid them.

An injury (“blessure”) can indeed also be described as a “lesion” (lésion) (the injury will most probably evolve into a lesion).

But a lesion caused by cancer or another disease, for example, is not an injury.

14

This has been a very interesting discussion and I thought presetning the rationale that SNOMED uses for the distinction between injury and lesion might be of some use:

A key point here is that injury and lesion are not semantically equivalent in SNOMED CT, even though some languages may use the same word for both. Trauma represents the inciting external event, injury represents the pathological process that occurs in the body in response to that event, and lesion represents the resulting structural abnormality. In other words, trauma is the cause, injury is the process, and lesion is the morphologic result. Lesions may arise from many causes, including trauma, infection, degeneration, or neoplasia, so lesion cannot be considered synonymous with injury.

It is also important to note that lesion is a general term for a morphologic abnormality and is typically used when the precise morphology or nature of the abnormality is not yet known. This is common in imaging, endoscopy, or initial clinical assessment, where an abnormality may be observed but cannot yet be more specifically characterized (for example, a “liver lesion” seen on CT). Once the nature of the abnormality is known, a more specific morphologic or disorder concept is usually more appropriate.

This distinction is important for modeling. Traumatic injury is most appropriately represented using relationships such as Due to = traumatic event, which identifies the external cause. SNOMED CT classifies injuries under the Disorder hierarchy because they represent pathological changes in body structure or function, whether explicitly modeled or implied. The traumatic event itself is modeled separately in the Event hierarchy as the external cause, while the injury represents the resulting pathological process and/or structural abnormality within the body. This approach allows the model to represent clearly the causal event, the pathological process, and the resulting morphology.

This also does not mean that all traumatic injury concepts must explicitly include the word “traumatic” in the FSN. The defining relationships are sufficient to represent the etiology, and FSN naming generally reflects common clinical usage rather than restating all defining characteristics. Many injury concepts are understood clinically to be traumatic unless otherwise specified, and representing this explicitly in the logical definition supports classification and analysis without requiring changes to established FSNs.

We recognize that in some languages there is no clear distinction between injury and lesion, and that this can create confusion or reduce clinical usefulness. This is a valid and important concern, and we appreciate the challenges of representing these distinctions consistently across different languages and clinical contexts. Lesion concepts remain important because they allow representation of structural abnormalities without asserting a specific cause or precise morphology, particularly in imaging and early diagnostic settings. We welcome continued discussion on how best to support national extensions in representing and using these concepts in ways that are clear and clinically meaningful for their users.

Overall, maintaining the distinction between trauma (event), injury (process), and lesion (morphology), and using lesion concepts appropriately as general descriptors when morphology is not yet known, helps ensure that SNOMED CT can accurately represent causation, biological processes, and structural abnormalities in a consistent and logically coherent way.

15

Thank you for this useful explanation, Jim. We understand the subtle theoretical semantic distinction between what is meant by lesion and by injury in SNOMED. However, if it’s taken many hours of debate for a group of specialized terminologists to understand it, it’d be very unlikely for everyday SNOMED users (for example, Belgian clinicians) to understand the difference in clinical settings.

For the time being, we will continue applying the Dutch approach of translating lesion as tissue anomaly and we will extend it to our French translations, in order to make a distinction, albeit a very artificial one, and to discourage the use of the “lesion” concepts. But ideally, national extensions should be allowed to selectively hide or inactivate certain international concepts (this would be useful for other hierarchies). That way we don’t have to spend resources coming up with terminological workarounds.

In Belgium, since we have three national languages, we maintain the English FSNs, which means that our clinicians will be able to view the word “lesion” and therefore select the lesion concept instead of the injury one, which rightfully contains the word “lésion” in the French synonyms. This can lead to the two concepts of a lesion-injury pairing being used interchangeably to register the same condition, which can cause inconsistencies in patient records, impact local/national healthcare statistics and hospital’s finances, and hinder international interoperability, which is more than just confusion or reduced clinical usefulness.