Musculo Skeletal Radiology

8/6/2019 Musculo Skeletal Radiology

1/57

Editors: Brant, William E.; Helms, Clyde A.

Title: Fundamentals of Diagnostic Radiology, 3rd Edition

Copyright 2007 Lippincott Will iams & Wilkins

> Table of Contents > Section X - Musculoskeletal Radiology > Chapter 41 - Benign Cystic

Bone Lesions

Chapter 41

Benign Cystic Bone Lesions

Clyde A. Helms

A benign, bubbly, cystic lesion of bone is one of the more common skeletal

lesions that a radiologist encounters. The differential diagnosis can be quitelengthy and is usually structured on how the lesion looks to the radiologist, usinghis or her experience as a guide. This method, called pattern identification,certainly has merit, but it can lead to a very long differential diagnosis and manyerroneous conclusions if not tempered with some logic.

In general, if a differential diagnosis wil l yield the correct diagnosis 95% of thetime, most would consider it a useful differential l ist; however, it would not be

appropriate to accept a 1-in-20 miss rate for fractures and dislocations. Ingeneral, the shorter the differential diagnosis l ist, the more helpful it is toclinicians and the easier it is to remember. A shorter differential l ist wil l usuallyhave a lower accuracy rate than a long list; however, many times the longer l ists

contain such rare entities that the accuracy does not really increasesubstantially. For most of the entities in bone radiology, a 95% accurate

differential is acceptable. If one wants to be more accurate than that, morediagnoses can simply be added to the l ist of differential possibil ities.

When the differential diagnosis is long, as in the differential for bubbly, cystic

lesions of bone, it can be difficult to recall all of the entities that should bementioned. A mnemonic can be helpful in recall ing long lists of information and

is recommended.

FEGNOMASHICFEGNOMASHICis a mnemonic that serves as a nice starting point for discussingpossibil ities that appear as benign, cystic lesions in bone. This mnemonic has

been in general use for many years. By itself, it is merely a long list14entitiesand it needs to be coupled with other criteria to shorten the l ist into amanageable form for each particular case. For instance, the age of the patient

will help add or eliminate many of the possibil ities. If multiple lesions arepresent, only half a dozen entities n eed to be discussed. Methods of narrowingthe differential are discussed later in this chapter.

The first step in approaching a benign, cystic bone lesion is to be certain it isreally benign. The criteria for differentiating benign lesions from malignant

Page 1 of 57Ovid: Fundamentals of Diagnostic Radiology

5/23/2011mk:@MSITStore:M:\medicalheaven_radiology2\Fundamentals%20of%20Diagnostic%20...


2/57

lesions are covered in Chapter 42. Once it is established that the lesion is truly abenign, cystic lesion, FEGNOMASHIC will enable a differential diagnosis that is at

least 95% accurate. Memorization of the 14 entities in this differential is easilydone (Table 41.1).

The next step after learning the names of all of the lesions is getting some idea

of each lesion's radiographic appearance. This is when experience becomes a

factor. For the medical student or first-year resident, it is difficult to go beyondsaying that they all look cystic, bubbly, and benign. The fourth-year residentshould have no trouble

differentiating between a unicameral bone cyst and a giant cell tumor because he

or she has seen examples of each many times before and knows theirappearance.

P.1064

TABLE 41.1 Discriminators for Benign Lytic Bone

LesionsMnemonic: FEGNOMASHIC

Letter Represents Characteristics

F Fibrous dysplasia No periosteal reaction

E Enchondroma 1. Calcification present (exceptin phalanges)

2. Pain less (no per iost i tis)

Eosinophi li c g ranu loma Younger than age 30

G Giant cel l tumor 1. Epiphyses closed2. Abuts the art icular surface (in

long bones)3. Wel l def ined w ith a

nonsclerotic margin (in longbones)

4. E ccen tri c

N Nonossifying fibroma 1. Younger than age 30

2. Pain less (no per iost i tis)3. Corti ca ll y based

O Osteoblastoma Mentioned when aneurysmal bone




3/57

After getting a feel for what each lesion looks l ike radiographically andovercoming the frustration that builds when one realizes that many of them look

alike, one should try to learn ways to differentiate each lesion from the others. Ihave developed a number of keys that I call discriminators, which help to

differentiate each lesion. These discriminators are 90% to 95% useful (I willmention when they are more or less accurate, in my experience) and are by nomeans intended to be absolutes or dogma. They are guidelines but have a highaccuracy rate.

Textbooks rarely state that a finding always o r never occurs. Theytemper descriptions with virtually always, invariably,usually, or characteristically. I have tried to pick out findings thatcome as close to always as I can, realizing that I wil l only beapproximately 95% accurate. That is good enough for most radiologists.

The following is only a brief description of each entity; more completedescriptions are readily available in any skel etal radiology text. What isemphasized here are the points that are unique for each entity, thereby enabling

cyst (ABC) is mentioned (especially

in the posterior elements of thespine)

M Metastatic diseasesand myeloma

Older than age 40

A Aneurysmal bone cyst 1. Expansile2. Younger than age 30

S Solitary bone cyst 1. Central

2. Younger than age 30

H Hyperparathyroidism(brown tumor)

Must have other evidence ofhyperparathyroidism

I Infection Always mention

C Chondroblastoma 1. Younger than age 302. Epiphyseal

Chondromyxoid No calcified matrix




4/57

differentiation from the others. Table 41.1 is a synopsis of these discriminators.

Fibrous DysplasiaFibrous dysplasia is a benign congenital process that can be seen in a patientof any age and can look like almost any pathologic process radiographically. It

can be wild looking, discretely lucent, patchy, sclerotic, expansile, multiple, and

many other descriptions. It is, therefore, difficult to look at a bubbly lytic lesionand unequivocally say it is or is not fibrous dysplasia. It would be better if theFEGNOMASHIC differential started on a positive note, say, with giant cell tumor

or chondroblastoma, for which there are some definite criteria. However, becausefibrous dysplasia is first on the l ist, we might as well deal with it.

How do you know whether to include or excl ude fibrous dysplasia if it can looklike almost anything? Experience is the best guideline. In other words, look in a

few texts and find as many different examples as possible; get a feeling for whatfibrous dysplasia looks l ike.

Fibrous dysplasia wil l not have periostitis associated with it; therefore, if

periostitis is present, one may safely exclude fibrous dysplasia. Fibrous dysplasia

FIGURE 41.1. Fibrous Dysplasia. This patient has polyostotic fibrous

dysplasia with diffuse involvement of the pelvis as well as the proximalfemurs.

P.1065




5/57

virtually never undergoes malignant degeneration and should not be a painfullesion unless there is a fracture. An occult fracture often occurs in long bones

with fibrous dysplasia; therefore, it is not unusual to have it present with painand no obvious fracture seen in a long bone. Pain in a flat bone, such as the ribs

or pelvis (nonweight-bearing bones), should not occur with fibrous dysplasia.

Fibrous dysplasia can be either monostotic (most commonly) or polyostotic and

has a predilection for the pelvis, proximal femur, ribs, and skull. When it ispresent in the pelvis, i t is invariably present in the ipsilateral proximal femur(Figs. 41.1, 41.2). I have seen only one case in which the pelvis w as involved

with fibrous dysplasia, and the proximal femur was spared. The proximal femur,however, may be affected alone, without involvement in the pelvis (Fig. 41.3).

Fibrous dysplasia often involves the ribs. It typically has an expansile, lyticappearance in the posterior ribs (Fig. 41.4) and a sclerotic appearance in the

anterior ribs.

The classic description of fibrous dysplasia is that it has a ground-glass or smokymatrix. This description confuses people as often as it helps them, and I do not

recommend using ground-glass appearance as a buzz word for fibrousdysplasia. Fibrous dysplasia is often purely lytic and becomes hazy or takes on a

ground-glass look as the matrix

calcifies (Fig. 41.5). It can go on to calcify significantly, and then it presents asa sclerotic lesion. Also, I often see lytic lesions with a pathologic diagnosis other

than fibrous dysplasia that have a distinct ground-glass appearance; therefore,the ground-glass qualifier can be misleading.

P.1066




6/57

FIGURE 41 . . Fibrous Dysplasia. This patient has polyostotic fibrousdysplasia with involvement of the right femur as well as the supraacetabular

portion of the ilium. When the pelvis is involved with fibrous dysplasia, the

ipsilateral femur on the affected side is invariably also involved.




7/57

FIGURE 41.3. Fibrous Dysplasia. This patient has a well-defined lytic

lesion with a hazy, ground-glass appearance in the neck of the right femur.

The pelvis was uninvolved. It is not unusual for monostotic fibrous dysplasiato involve the proximal femur and spare the pelvis.




8/57

FIGURE 41.4. Fibrous Dysplasia. When fibrous dysplasia affects the ribs,

the posterior ribs often demonstrate a lytic expansile appearance, as in this

example. When the anterior ribs are involved, they are most often sclerotic

in appearance. Note also the involvement of the thoracic spine.




9/57

AdamantinomaWhen a lesion i s encountered in the tibia that resembles fibrous dysplasia, an

adamantinoma should also be considered. An adamantinoma is a malignanttumor that radiographically and histologically resembles fibrous dysplasia (Fig.

41.6). It occurs almost exclusively in the tibia and the jaw (for unknown

reasons) and is rare. Because it is rare, one may choose not to include it in the

differentiala misdiagnosis will not occur more than once or twice in a lifetime.

McCune Albrig t SyndromePolyostotic fibrous dysplasia occasionally occurs in association with caf-au-lait

FIGURE 41 . . Fibrous Dysplasia. Polyostotic fibrous dysplasia is seen in

the radius in this child. Parts of this lesion have a hazy, ground-glass

appearance, whereas others are more l ytic appearing. A hazy, ground-glass

appearance is often present in fibrous dysplasia, but just as often, the

appearance can be purely lytic or even sclerotic.




10/57

spots on the skin (dark-pigmented, frecklelike lesions) and precocious puberty.

This complex is called McCune-Albright syndrome. The bony lesions in this

syndrome, and even in the simple polyostotic form, often occur

unilaterallythat is, throughout one half of the body. This

does not happen often enough to be of any diagnostic use in differentiating

fibrous dysplasia from other lesions.

The presence of multiple lesions of fibrous dysplasia in the jaw has been termed

c er ism This is from the physical appearance of the child with puffed-out

cheeks having an angelic look. The jaw lesions of cherubism regress in

P.1067

FIGURE 41 . . Adamantinoma. This mixed lytic and sclerotic process in the

midshaft of the tibia is characteristic for fibrous dysplasia. An

adamantinoma has an identical appearance and should be considered in any

tibial lesion that resembles fibrous dysplasia. Biopsy showed this to be an

adamantinoma.




11/57

adulthood.

Discriminator

No periosteal reaction.

Enchondroma and Eosinophilic GranulomaEnchondromas occur in any bone formed from cartilage and may be central,

eccentric, expansile, or nonexpansile. They invariably contain calcified chondroid

matrix, except when in the phalanges. An enchondroma is the most common

benign cystic lesion in the phalanges (Fig. 41.7). If a cystic lesion is present

without calcified chondroid matrix anywhere except in the phalanges, I do not

include enchondroma in my differential.

Often it is difficult to differentiate between an enchondroma and a bone infarct.

An infarct usually has a well-defined, densely sclerotic, serpiginous border (Fig.

41.8), whereas an enchondroma does not (Fig. 41.9). An enchondroma often

causes endosteal scalloping, whereas a bone infarct will not. Although these

criteria are helpful in separating an infarct from an enchondroma, they are not

foolproof.




12/57

It is difficult, if not impossible, to differentiate an enchondroma from a

chondrosarcoma. Clinical findings (primarily pain) serve as a better indicatorthan radiographic findings, and indeed pain in an apparent enchondroma should

warrant surgical investigation. Periostitis should not be seen in an enchondroma

either. Trying to differentiate histologically an enchondroma from a

chondrosarcoma is also difficult, if not impossible, at times. Biopsy of an

apparent enchondroma should not be performed routinely for histologic

differentiation.

Multiple enchondromas occur on occasion; this condition has been termed

FIGURE 41 . . Enc ondroma. A lytic lesion in the phalanges is most

commonly an enchondroma. This is the only location in the skeleton where

an enchondroma does not contain calcified chondroid matrix. These most

often present with pathologic fractures, as in this example.




13/57

Ollier disease (Fig. 41.10). It is not

hereditary and does not have an increased rate of malignant degeneration. The

presence of multiple enchondromas associated with soft tissue hemangiomas is

known as Maffucci syndrome (Fig. 41.11). This syndrome also is not hereditary;

however, it does have an increased incidence of malignant degeneration of theenchondromas.

P.1068

P.1069

FIGURE 41 . . Bone Infarct. These lytic lesions in the distal femurs with

calcified, serpiginous borders are typical for bone infarcts. Occasionally, the

differential between a bone infarct and an enchondroma can be difficult on

plain films; however, in this example, infarcts are easily diagnosed.




14/57

FIGURE 41 . . Enc ondroma. This lesion in the distal right femur shows

the stippled, punctate calcification typical for chondroid matrix seen in an

enchondroma.




15/57

FIGURE 41.1 . Ollier Disease. Multiple enchondromas are present

throughout the hand. This is a typical example of Ollier disease.




16/57

Discriminators1. Must have calcification (except in phalanges). 2. No periostitis or pain.

Eosinophilic granuloma (EG) is a form of histiocytosis X, the other forms being

Letterer-Siwe disease and Hand-Schller-Christian disease. Although theseforms may be merely different phases of the same disease, most investigators

categorize them separately. The bony manifestations of all three disorders are

similar and are discussed in this review simply as EG.

Unfortunately for radiologists, EG has many appearances (1). It can be lytic or

sclerotic, it may be well defined or ill defined, it might or might not have a

sclerotic border, and it might or might not elicit a periosteal response. The

periostitis, when present, is typically benign in appearance (thick, uniform,

FIGURE 41.11. Maffucci Syndrome. Multiple enchondromas associatedwith phleboliths are present in the phalanges. This combination of findings

invariably represents hemangiomas and enchondromas in Maffucci

syndrome.




17/57

wavy) but can be lamellated or amorphous. EG can mimic Ewing sarcoma and

present as a permeative (multiple small holes) lesion.

How, then, can one distinguish EG from any of the other lytic lesions in this

differential? Remember that it is difficult to exclude EG from almost any

differential of a bony lesion, be it benign or malignant. EG occurs almost

exclusively in patients under 30 years (usually


18/57

EG might or might not have a soft tissue mass associated, so the presence or

absence of a soft tissue mass will not help in the differential diagnosis. I know of

no entity in which the presence or absence of an associated soft tissue mass will

warrant inclusion or exclusion of the process from a differential. It is important

to note the presence of a soft tissue mass (or its absence), but this will do little

to narrow the differential diagnosis.

Most radiologists are inept at evaluating the soft t issues because they aredifficult to see, and CT and MR have made

it unnecessary in most cases to rely on plain films alone for the soft tissues.

Fortunately, in most cases, the presence or absence of a soft tissue mass will not

alter the differential diagnosis. The treating physician will undoubtedly want to

know whether the soft tissues are involved and to what extent; this can be

satisfactorily demonstrated with MR.

P.1070




19/57

EG occasionally has a bony sequestrum (Fig. 41.14). Only a few other entities

have been described that on occasion have bony sequestra: osteomyelitis,

lymphoma, and fibrosarcoma; therefore, when a sequestrum is identified, EG,

osteomyelitis, lymphoma, and fibrosarcoma should be considered. As discussed

in Chapter 47, an osteoid osteoma will often give an appearance of a

sequestration when the nidus is partially calcified. Clinically, EG might or might

not be associated with pain; therefore, clinical history is noncontributory for the

most part.

DiscriminatorMust be under age 30 years.

Giant Cell umor

FIGURE 41.13. Eosinop i lic G ranuloma EG . Well-defined lytic lesions

are present throughout the pelvis in this 24-year-old patient. In addition tothe lesion around the right hip, a lesion is seen a t the right sacroiliac joint.

Biopsy showed this to be EG.




20/57

Giant cell tumor is an uncommon, somewhat controversial lesion with several

schools of thought as to its radiographic appearance. I subscribe to the most

widely used approach, as do the majority of radiologists and pathologists (2).

First, it is important to realize that one is unable to tell whether a giant cell

tumor is benign or malignant, regardless of its radiographic appearance.

Histologically, a giant cell tumor cannot be divided into either a benign or a

malignant category. Most surgeons curettage and pack the lesions and consider

them benign unless they recur. Even then, they can still be benign and recur a

second or third time. About 15% of giant cell tumors are thought to be malignant

on the basis of their recurrence rate. When malignant, they can metastasize to

the lungs, but they do so infrequently.

Four classic radiographic criteria for diagnosing giant cell tumors exist. If any of

these criteria are not met when looking at a lesion, giant cell tumor can beeliminated from the differential diagnosis.

1. Giant cell tumor occurs only in patients with closed epiphyses; this is valid

at least 98% to 99% of the time and is extremely useful. I do not entertain

the diagnosis of giant cell tumor in a patient with open epiphyses.

2. The lesion must be epiphyseal and abut the articular surface (Fig. 41.15).

FIGURE 41.14. Eosinop i lic G ranuloma EG . This well-defined lytic

lesion contains a bony sequestrum (arro ), which is typical for osteomyelitis

or EG. Biopsy revealed this to be EG.




21/57

There is disagreement as to whether giant cell tumors begin in the

epiphyses or metaphyses or from the physeal plate itself; however, except

for rare cases, when radiologists see the lesions, they are epiphyseal and

are flush against the articular surface. The metaphysis also has some of the

tumor in it because the lesions are generally very large. When one sees a

giant cell tumor, it will be epiphyseal. Perhaps more importantly, it should

be flush against the articular surface of the joint. This occurs in 98% to99% of giant cell tumors; therefore, if I have a lesion that is

separated from the articular surface by a definite margin of normal bone, I

will not include giant cell tumor in the diagnosis. This rule does not apply in

flat bones, such as i n the pelvis or in the apophyses (Fig. 41.16), which

have no articular surfaces.

P.1071

FIGURE 41.1 . Giant Cell Tumor. A well-defined lytic lesion without

a sclerotic margin is seen abutting the articular surface of the distal

femur in a patient who has closed epiphyses. These are all




22/57

3. Giant cell tumors are said to be eccentrically located in the bone, as

opposed to being centrally placed in the medullary cavity. When a bonylesion is quite large, it can be difficult to tell whether it is central or

eccentric. I do not find this to be a terribly useful description, but it is one

of the classic rules of a giant cell tumor.

4. The lesion must have a sharply def ined zone of transition (border) that is

not sclerotic. This is a very helpful finding in giant cell tumor. The only

places this does not apply is in flat bones, such as the pelvis (Fig. 41.17)

and the calcaneus.

It is important to realize that the four criteria for a giant cell tumor apply only to

giant cell tumors and not to any other lesion. For instance, I know of no other

lesion that depends on whether the epiphyses are open or closed. No other lesion

in any of my lists uses as a diagnostic factor whether the zone of transition is

sclerotic or not (many lesions, such as nonossifying fibromas, will usually have a

sclerotic margin, but it does not occur often enough to include as a

discriminator). No other lesion must always abut the articular surface, and no

other lesion has the classic

description of being eccentrically placed (although several lesions, including

nonossifying fibroma and chondromyxoid fibroma, are eccentric most of the

time).

characteristics of a giant cell tumor.

P.1072




23/57

FIGURE 41.1 . Giant Cell Tumor. This well-defined lytic lesion that does

not have a sclerotic margin completely involves the greater trochanter. The

apophyses have the same differential diagnosis as lesions in the epiphyses,

which makes giant cell tumor a strong possibility in this example. Biopsy

showed this to be a giant cell tumor.




24/57

FIGURE 41.1 . Giant Cell Tumor. A large, well-defined lytic lesion in the

iliac wing is seen, which does contain a sclerotic margin and does not

appear to abut any articular surface. The pelvis is a good location for giant

cell tumor, which this proved to be at biopsy. The usual rules for giant cell

tumors such as presence of a nonsclerotic margin do not apply in flat bones.

FIGURE 41.1 . Fibrous Cortical Defect. A well-defined lytic lesion is seen

in the medial metaphysis of this tibia (arro s), which is typical for a fibrous

cortical defect.




25/57

FIGURE 41.1 . Nonossifying Fibroma. A large, well-defined lytic lesion,

which is slightly expansile with scalloped sclerotic margins, is seen in the

distal tibia in this young patient. This is a characteristic appearance of a

nonossifying fibroma. The examination was obtained for a sprained an kle

and not for this asymptomatic lesion.




26/57

Although these four criteria apply well for giant cell tumor, they do not apply at

all for any other lesions. Residents have a tendency to apply these criteria to

every lytic lesion encountered for the simple reason that they have learned the

four criteria.

Once one of the criteria is violated, the remainder do not even have to be used

to eliminate a giant cell t umor. For instance, if a lytic lesion is found in the mid-

diaphysis of a bone, giant cell tumor can be excluded. There is no need to check

further to see whether it is eccentric, whether it has a nonsclerotic margin, or

FIGURE 41. . Nonossifying Fibroma. A well-defined, expansile lytic

lesion in the distal f ibula is noted in this asymptomatic patient, which is

characteristic for a nonossifying fibroma.




27/57

whether the epiphyses are closed.

Again, these rules will be greater than 95% effective and, in my experience,

close to 99% effective. It should be noted that these criteria only apply to giant

cell tumors of long bones. They would not work, for instance, in the pelvis or the

calcaneus, two locations where giant cel l tumors often occur. If one or two cases

are found that do not fit the criteria, another pathologist should review the

slides.

Many pathologists refer to aneurysmal bone cysts as giant cell tumors; hence,

they have giant cell tumors that do not obey any of the criteria. These

pathologists may be correct, but they are not in the mainstream of what most

people use for giant cell tumor criteria, both radiographically and histologically.

Discriminators1. Epiphyses must be closed. 2. Must abut the articular surface. 3. Must be well

defined with a nonsclerotic margin. 4. Must be eccentric.

onossifying Fibroma onossifying f ibroma (NOF) is probably the most common bone lesion

encountered by radiologists. It reportedly occurs in up to 20% of children and

usually spontaneously regresses, so as to be seen only rarely after the age of

30. Fibrous cortical defect is a c ommon synonym, although some people

differentiate the two lesions on the basis of size, with a fibrous cortical defect

being smaller than 2 cm in length (Fig. 41.18) and an NOF being larger than 2

cm (Fig. 41.19). Histologically, these lesions are identical; therefore, it seems

appropriate to refer to them all as NOFs rather than to subdivide them by their

size.

P.1073




28/57

NOFs are benign, asymptomatic lesions that typically occur in the metaphysis of

a long bone, emanating from the cortex. They classically have a thin, sclerotic

border that is scalloped and slightly expansile (Fig. 41.20); however, this is a

general description that probably applies to o nly 75% of the lesions and could

equally apply to most of the lesions in FEGNOMASHIC. They do not have to haveexpansion or a scalloped or sclerotic border and are not limited to the

metaphyses. Then how are they best recognized? The best way is to familiarize

oneself with their general appearance by looking at examples in textbooks. That

can be done in 15 minutes. It is important to recognize these lesions because

they are what I call dont touch lesions (see Chapter 46); that is, the

radiologist's diagnosis should be the final word and thereby supplant a biopsy.

These lesions are so characteristic that no differential diagnosis should be

FIGURE 41. 1 . Nonossifying Fibroma. A. A well-defined, lytic lesion that

is minimally expansile is seen in the distal ti bia in this child who was

examined for a sprained ankle. B. A CT examination showed apparentcortical destruction (arro ), which was believed to be suggestive of an

aggressive lesion. Biopsy showed this to be a nonossifying fibroma. Both CT

and MR will often show apparent cortical destruction, which is merely

cortical replacement by benign fibrous tissue.




29/57

entertained, although a few entities can indeed occasionally simulate them.

If a CT or MR is obtained of an NOF, there will often appear to be interruption of

the cortex, which can be misinterpreted as cortical destruction (Fig. 41.21). This

merely represents cortical replacement by benign fibrous tissue and should not

warrant further investigation.

P.1074

FIGURE 41. . Healing Nonossifying Fibroma. A predominantly sclerotic

lesion, which is minimally expansile and well defined, is seen in the proximal

humerus in this child who is asymptomatic. This is a typical appearance of a

disappearing or healing nonossifying fibroma. With time, this l esion will meltinto the normal bone and essentially disappear.




30/57

FIGURE 41. 3 . Nonossifying Fibroma. This large, well-defined lytic lesion

with faint sclerotic margins is seen in the distal femur. It has a very typical

appearance for a giant cell tumor; however, it has sclerotic margins and

does not abut the arti cular surface. The lesion underwent biopsy and was

found to be a n onossifying fibroma.




31/57

If the patient is older than 30 years of age, NOF should not be included in the

differential diagnosis. NOFs must be as ymptomatic and exhibit no periostitis,

unless there is an antecedent history of trauma. They routinely heal with

sclerosis and eventually disappear (Fig. 41.22), usually around the ages of 20 to

30 years. During this healing period, they can appear hot on a radionuclide bone

scan because there is osteoblastic activity. These lesions can occasionally get

quite large (Fig. 41.23); therefore, growth or change in size should not alter the

diagnosis. They are most commonly seen about the knee but can occur in any

long bone. Occasionally, multiple NOFs are seen about the knee, each of which is

characteristic in appearance.

Discriminators1. Must be younger than age 30 years. 2. No periostitis or pain. 3. Cortically

based.

steoblastoma s teoblastomas are rare lesions that co uld justif iably be excluded from this

differential without the fear of missing a diagnosis more than once in a lifetime.

Why, then, include them? The mnemonic FEGNOMASHIC would not have nearly

the same ring without the extra vowel, so osteoblastoma remains.

Osteoblastomas have two appearances: (1) They look like large osteoid osteomas

and are often called i ant osteoid osteomas Because osteoid osteomas are

FIGURE 41. 4 . Osteoblastoma. A. A lytic expansile lesion involving the

right T12 pedicle (arro ) and transverse process is seen on thisanteroposterior plain film. B. The lesion is seen on CT to extend into the

vertebral body. It has intact cortices and contains some calcif ied matrix.

This is a classic example of an osteoblastoma of the spine.

P.1075




32/57

sclerotic lesions and do not resemble bubbly lytic lesions, this is not the type of

osteoblastoma we are concerned with in this differential. (2) They simulate

aneurysmal bone cysts (ABCs). They are expansile, often having a soap-

bubble appearance. If an ABC is being considered, so should an

osteoblastoma. Osteoblastomas commonly occur in the posterior elements of the

vertebral bodies, and about half of the cases demonstrate speckled calcif ications

(Fig. 41.24). A classic radiology differential is that of an expansile lytic lesion ofthe posterior elements of the spine, which includes osteoblastoma, ABC, and

tuberculosis.

DiscriminatorMentioned when ABC is mentioned (especially in the posterior elements of the

spine).

Metastatic Disease and MyelomaMetastatic disease should be considered for any lytic lesionbenign or

aggressive in appearancein a patient over 40 years of age. Metastatic diseasecan appear perfectly benign radiographically (Fig. 41.25), so it i s not valid to

say, Because this lesion looks benign, it should not be a metastasis. Most

metastatic disease has an aggressive appearance and will not be in the

FEGNOMASHIC differential, but a significant number of metastases appear

benign. In fact, metastases can have any radiographic appearance; therefore,

any bone lesion in a patient older than the age of 40 should have metastatic

disease as a consideration, unless trauma or arthritis is the primary concern.




33/57

For statistical purposes, I do not mention metastatic disease in a pati ent younger

than age 40. I will be correct more than 99% of the time using 40 as a cutoff

age. Otherwise, metastatic diseases would have to be mentioned in every single

case of a lytic lesion, and I prefer to limit the list of differential possibilities. I

am not claiming that met astatic disease does not occur in patients younger thanage 40only that I consider it acceptable to miss it (unless given a history of a

known primary neoplasm).

MyelomaAlthough myeloma most commonly presents as a diffuse permeative process in

the skeleton (Fig. 41.26), it can present as either a solitary lesion (Fig. 41.27)

or as multiple lytic lesions. Bubbly, lytic bone lesions of myeloma are more

FIGURE 41. . Metastatic Disease. A well-defined lytic lesion is seen inthe proximal femur in this 50-year-old patient who had pain associated with

this lesion. Biopsy showed this to be a renal metastasis. A significant

number of metastatic lesions can have a completely benign appearance, as

in this example.




34/57

correctly called p a smac t omas I mention plasmacytoma separately from

metastatic disease because it can occur in a slightly younger population (age

greater than 35 years is my cutoff) and can precede clinical or hematologic

evidence of myeloma by 3 to 5 years. In general, there is no harm in lumping all

metastatic disease, including myeloma, into one group and using greater than

age 40 as the limiting factor.

Virtually any metastatic process can present as a lytic, benign-appearing lesion;

therefore, it serves no purpose to

try to guess the source of the metastatic disease from its appearance. In

general, lytic expansile metastatic diseases tend to come from thyroid and renal

tumors (Fig. 41.28). The only metastatic lesion that is said to always be lytic is

renal cell carcinoma.

Discriminator

Must be older than age 40 years.

P.1076

FIGURE 41. . Multiple Myeloma. A. A diffuse permeative pattern is

present throughout the femur in this patient wit h multiple myeloma. B. A

lateral skull f ilm shows a typical presentation of multiple myeloma in the

skull with multiple small holes throughout the calvaria, which are well

defined.




35/57

FIGURE 41. . l asmacytoma. A large, well-defined lytic lesion is seen in

the left ilium (arro s) in this patient with multiple myeloma. This is a

common location for a plasmacytoma. Like metastases, plasmacytomas often

have a completely benign appearance.




36/57

neurysmal one Cyst neurysmal bone cysts (ABCs) are the only lesions I know of that are named

for their radiographic appearance. They are virtually always aneurysmal or

expansile (Figs. 41.29, 41.30). Rarely, an ABC will present before it is expansile,

but that is unusual enough not to worry about. Aneurysmal bone cysts occur

primarily in patients who are younger than age 30, although occasionally one will

be encountered in older patients. I use bony expansion and age of less than 30

years as fairly rigid guidelines and seldom miss the diagnosis of ABC. They often

have fluidfluid levels on CT or MR (Fig. 41.31), although this is a nonspecific

FIGURE 41. . Metastatic Disease. An expansile lesion with a soap-

bubble appearance is present in the proximal radius in a patient with renal

cell carcinoma. An expansile lytic lesion is a common finding with renal or

thyroid metastatic disease.

P.1077




37/57

f inding, as many other lesions can have fluidfluid levels.

ABCs are, like giant cell tumors, somewhat controversial. There are apparently

two types of ABCs: a primary type and a secondary type. The secondary type

occurs in conjunction with another lesion or from trauma, whereas a primary ABC

has no known cause or association with other lesions. Secondary ABCs have been

said to occur with giant cell tumors, osteosarcomas, and almost any other lesion.

I have seen dozens of ABCs and have seen only a few in association with another

lesion. As to occurring after trauma, I do not understand why they would be age-

limited if trauma were causative. Also, malignant tumors

were once thought to occur after trauma because of the frequent association of a

history of antecedent trauma with malignant bone tumors. This is not seriously

considered today and is thought to be coincidental. I suspect that ABCs and

trauma are also coincidental, but this is mere speculation.

P.1078

FIGURE 41. . Aneurysmal Bone Cyst. An expansile lytic lesion is present

in the distal femur in this 24-year-old patient who presents with pain. This

is a fairly typical appearance for an aneurysmal bone cyst.




38/57

FIGURE 41.3 . Aneurysmal Bone Cyst. A well-defined expansile lesion is

seen in the midshaft of the ulna in a child who presented with pain in this

region. This is a characteristic appearance for an aneurysmal bone cyst.




39/57

ABCs typically present because of pain. They can occur anywhere in the skeleton,

and there is no location that would make them more highly ranked in the

differential diagnosis. As with osteoblastoma, they often o ccur in the posteriorelements of the spine.

FIGURE 41.31. Aneurysmal Bone Cyst. An axial T2WI through a thoracic

vertebral body shows an expansile lesion involving the posterior elements

that has several fluidfluid levels (arro s). This is a typical appearance

for an aneurysmal bone cyst.




40/57

FIGURE 41.3 . Solitary Bone Cyst. A well-defined lytic lesion is present in

the proximal humerus in this child who suffered a fracture through the

lesion. The location and central appearance, as well as the age of the

patient, are characteristic for a solitary bone cyst. A piece of cortical bone

has broken off and descended through the serous fluid contained within the

lesion and can be seen in the dependent portion of the lesion (arro ) as a

fallen fragment sign. A fallen fragment sign is said to be pathognomonic for

a unicameral bone cyst.




41/57

Discriminators1. Must be expansile. 2. Must be younger than age 30 years.

olitary one Cyst o li tary bone cysts are also called simp e one c sts or nicamera one c sts

They are not necessarily unicameral (one compartment), however. This is the

only lesion in FEGNOMASHIC that is always central in location. Many of the other

lesions may be central, but a solitary bone cyst can be excluded if it is not. It is

one of the few lesions that does not occur most commonly around the knees. Two

thirds to three fourths of these lesions occur in the proximal humerus (Fig.

41.32) and proximal femur (Fig. 41.33). Application of this rule alone is not that

helpful, or one third to one fourth of lesions would be missed.

Solitary bone cysts are usually asymptomatic unless fractured, which is a

common occurrence. Even when pathologic fractures occur, they rarely form

periostitis. A classic radiographic finding for a solitary bone cyst is the fallen

fragment sign (see Fig. 41.32). This occurs when a piece of cortex breaks offafter a fracture in a so litary bone cyst, and the piece of cortical bone sinks to

the gravity-dependent portion of the lesion. This has not been

described in any other lesion and indicates a fluid-filled cystic lesion, rather than

a lesion fi l led with m atrix.

P.1079




42/57

FIGURE 41.33. Solitary Bone Cyst. A well-defined lytic lesion, which is

central in location, is seen in the proximal femur in this child. This is

characteristic for a solitary bone cyst.




43/57

FIGURE 41.34. Solitary Bone Cyst. A well-defined lytic lesion is seen in

the calcaneus abutting the inferior surface, which is typical in location and

appearance for a solitary bone cyst. A solitary bone cyst in the calcaneus

occurs almost exclusively in this location and is not subject to pathologic

fracture as readily as when one occurs in the proximal femur and humerus.




44/57

Solitary bone cysts occur almost exclusively in young patients (under age 30).

Although long bones are most commonly involved, solitary bone cysts have been

described in almost every bone in the body. They begin at the physeal plate in

long bones and grow into the shaft of the bone; therefore, they are not

epiphyseal lesions. They can, however, extend up into an epiphysis after the

plate closes, but this is unusual. A fairly common location is in the calcaneus,

where they have a characteristic location adjacent to the inferior surface of the

calcaneus (Fig. 41.34).

Discriminators1. Must be central. 2. Must be younger than age 30 years. 3. No periostitis.

yperparathyroidism ro n umors ro n tumors of hyperparathyroidism (HPT) can have almost any appearance,

from a purely lytic lesion (Fig. 41.35) to a sclerotic process. Generally, when the

patient's HPT is treated, the brown tumor undergoes sclerosis and will eventually

disappear. If a brown tumor is going to be considered in the differential

diagnosis, additional radiographic findings of HPT should be seen. Subperiosteal

bone resorption is pathognomonic for HPT and should be

searched for in the phalanges (particularly in the radial aspect of the middle

phalanges) (Fig. 41.35), distal clavicles (resorption), medial aspect of the

proximal tibias, and sacroiliac joints. If the physes are open, they should have a

frayed, ragged appearance, as in rickets, owing to the effect of parathormone.

Osteoporosis or osteosclerosis might suggest that renal osteodystrophy with

secondary HPT is present, but subperiosteal resorption must be present, or

brown tumor can be safely excluded from the differential.

FIGURE 41.3 . Bro n Tum or. A. An expansile lytic lesion is seen in the

fifth metacarpal (arro s), and a second, smaller lytic lesion is seen in the

proximal portion of the fourth proximal phalanx. B. This patient is noted to

have subperiosteal bone resorption, best seen in the radial aspect of the

middle phalanges (arro s) as indistinct, interrupted cortex. This makes the

diagnosis of hyperparathyroidism with multiple brown tumors most likely.

P.1080




45/57

Most authorities believe that brown tumors occur most commonly in primary

HPT; however, because we see so many more patients with secondary HPT, more

brown tumors are seen in patients with secondary rather than primary HPT.

DiscriminatorMust have other evidence of HPT.

Infection

InfectionUnfortunately, there is no reliable way radiographically to exclude a focus of

osteomyelitis. It has a protean radiographic appearance and can occur at any

location and in a patient of any age. It might or might not be expansile, have a

sclerotic or nonsclerotic border, or have associated periostitis (3). Therefore,

infection will be in almost every differential diagnosis of a lytic lesion, which is

acceptable, as it is one of the most common lesions encountered. Soft tissue

findings such as obliteration of adjacent fat planes are notoriously unreliable and

even misleading, because tumors and EG can do the same thing.

When osteomyelitis occurs near a joint, if the articular surface is abutted,

invariably the joint will be involved and show cartilage loss, an effusion (Fig.

FIGURE 41.3 . Osteomyelitis. A. A plain film of the proximal humerus in

this child with shoulder pain reveals a well-defined lytic lesion in the medial

metaphysis. B. T2WI of the humerus shows the lesion to have high signal

and an associated joint effusion. The probable site of connection to the joint

can be seen (arro ), which likely represents a draining abscess. Aspiration

of the joint fluid revealed pus. This is a large focus of osteomyelitis or

Brodie abscess.




46/57


47/57

FIGURE 41.3 . Osteomyelitis. A. A lytic lesion is present in the proximal

humerus, which has some associated periostitis laterally. B. CT scan through

this area reveals a lytic lesion that contains a calcific density within (arro ),

which is a bony sequestrum. This is an area of osteomyelitis with a bony

sequestration.




48/57

The differential diagnosis of a lytic lesion in the epiphysis of a patient under 30

years of age is simple: (1) infection (most common), (2) chondroblastoma, and

(3) giant cell tumor (which has its own diagnostic criteria, so it can usually be

definitely ruled out or in). This is an old, classic differential and probably

encompasses 98% of epiphyseal lesions.

A caveat on epiphyseal lesions is to always consider the possibility of a

subchondral cyst or geode (Fig. 41.39), which has been described in four disease

processes: (1) degenerative joint disease (must have joint space narrowing,

sclerosis, and osteophytes); (2) rheumatoid arthritis; (3) calcium pyrophosphate

dihydrate crystal disposition disease or pseudogout; and (4) avascular necrosis.

The clinician must be certain no joint pathology that might indicate one of these

processes is present, or an unnecessary biopsy of a geode might be performed

on the basis of the differential of an epiphyseal lesion.

Apophyses are identical to epiphyses as far as the differential diagnosis of lytic

lesions, with the exception of geodes, which only occur adjacent to articular

surfaces. The carpal bones, the tarsal bones, and the patella have

FIGURE 41.3 . C ondroblastoma. A plain film in this young patient shows

a well-defined lytic lesion in the greater tuberosity of the humerus. Biopsy

showed this to be a chondroblastoma.




49/57

a tendency to behave like epiphyses in their differential diagnosis of lesions.

Therefore, a lytic lesion in these areas has a similar differential diagnosis as an

epiphyseal lesion.

Discriminator1. Must be younger than age 30. 2. Must be epiphyseal.

Chondromy oid FibromaChondromy oid fibroma, like osteoblastoma, is such a rare lesion that failure

to mention it is probably not going to result in missing more than one in a

lifetime. Why include it, then? I recommend not including it, but it is part of the

classic FEGNOMASHIC differential. If it is mentioned, at least know what it looks

like. Basically, chondromyxoid fibromas resemble NOFs. Unlike NOFs, however,

they can be seen in a patient of any age. Chondromyxoid fibromas often extend

P.1082

FIGURE 41.3 . Geode. A large, well-defined lytic lesion in the proximal

humerus is present, which is associated with marked degenerative disease

of the glenohumeral joint. When definite degenerative joint disease is

present and associated with a lytic lesion, the lytic lesion should be

considered to be a geode. A biopsy was performed, which confirmed this to

be a geode, or subchondral cyst; however, the biopsy could have been

avoided.




50/57

into the epiphyses (Fig. 41.40), whereas NOFs rarely do. Also, they can present

with pain, which will not occur with an NOF. They have been reported to progress

from a benign process to an aggressive and even malignant lesion, but this is

extremely rare. Although chondromyxoid fibromas are cartilaginous lesions,

calcified cartilage matrix is virtually never seen radiographically.

Discriminator1. Mention when an NOF is mentioned. 2. No calcified matrix.

FIGURE 41.4 . C ondromy o id F ib roma. A well-defined lytic lesion in the

distal tibia that extends slightly into the epiphysis is noted on thisanteroposterior plain film. A nonossifying fibroma could certainly have this

appearance; however, this underwent biopsy and was found to be a

chondromyxoid fibroma. Chondromyxoid fibromas often extend into the

epiphysis, as in this example, whereas nonossifying fibromas usually do not.




51/57

S MMA That, in essence, is the differential diagnosis for a benign cystic lesion of bone.

It is probably 98% accurate, which is good enough for most radiologists. To

increase the accuracy to 99%, it would be necessary to add many uncommon or

rare lesions, and the whole process would become too confusing for most

radiologists to learn and

apply. If there is a favorite lesion that is not on this list, by all means add it.

Likewise, if the list is already too cumbersome, forget about osteoblastoma and

chondromyxoid fibroma. I am unable to make it much simpler than that and still

be reasonably accurate.

P.1083

TABLE 41. Lesions in atients ounger T an ears of Age

Eosinophilic granuloma

Aneurysmal bone cyst

Nonossifying fibroma

Chondroblastoma

Solitary bone cyst




52/57

Some of the l esions I have purposefully omitted are intraosseous ganglion,

pseudotumor of hemophilia, hemangioendothelioma, ossifying fibroma,

intraosseous lipoma, glomus tumor, neurofibroma, plasma cell granuloma, and

schwannoma. Others could be added to this list, of course, but are best left to

the pathologistnot the radiologistfor the diagnosis.

There are several features that are somewhat useful in separating the various

lesions in FEGNOMASHIC. For instance, if the patient is younger than the age of

30 years, be sure to consider EG c ondro astoma NOF so itar one c st , and

A C (Table 41.2). If the patient is over 30 years of age, those five lesions can be

excluded. Note that this is not a differential diagnosis for lesions in patients

under age 30; it simply means these entities should not be mentioned in older

patients. For those younger than age 30, other lesions such as fibrous dysplasia

and infection must also be mentioned.

There are a few lytic lesions that have no good discriminators other than age

and, therefore, must be mentioned routinely. I call these lesions

automatics because one should automatically mention them regardless of

the location or appearance of the lesion. Infection and EG must be mentioned for

those younger than age 30, whereas metastatic disease and infection must be

included in any differential in a patient older than age 40 (Table 41.3). These

lesions have a protean radiographic appearance and should be mentioned not

only in the benign cystic differential but also for an aggressive lesion.

TABLE 41. Automatics

Younger than age 30

InfectionEosinophilic granuloma

Older than age 40

Infection

Metastatic disease and myeloma

TABLE 41.4 Lesions T at Ha e No ain or eriostitis




53/57

If periostitis or pain is present (assuming no trauma, which can be a foolhardy

assumption), you can exclude fi ro s d sp asia so itar one c st NOF , and

enc ondroma (Table 41.4). If the lesion i s epiphyseal, the differential is

infection iant ce t mor c ondro astoma (and do not forget eodes) (Table

41.5). If the patient is over 40 years of age, add metastatic disease and

m e oma and remove c ondro astoma from the epiphyseal list.

The epiphyseal differential tends to apply also to the tarsal bones (especially the

calcaneus), the carpal bones, and the patella. In the calcaneus, a unicameral

bone cyst should also be considered and has a characteristic appearance and

location (see Fig. 41.34). Apophyses are epiphyseal equivalents and have

the same differential as epiphyses. The difference between an epiphysis and an

apophysis is that epiphyses contribute to the length of a bone, whereas

apophyses serve as ligamentous attachments.

Fibrous dysplasia

Enchondroma

Nonossifying fibroma

Solitary bone cyst

TABLE 41. Epip yseal Lesions

Infection

Giant cell tumor

Chondroblastoma

Geode




54/57

A classic differential for benign, cystic rib lesions is the mnemonic FAME, in

which F = f i ro s d sp asia, A = A C , M = metastatic diseases and m e oma,

and E enc ondroma and EG (Table 41.6). If multiple lytic l esions are present,

FEEMHI is a useful mnemonic of the lesions in FEGNOMASHIC that can be

multiple: F = fi ro s d sp asia, E enc ondroma, E EG, M = metastatic

disease and m e oma, H perparat roidism (brown tumor), and I i nfection

(Table 41.7).

A few findings that just do not seem to narrow the differential diagnosis are

presence or absence of a s oft tissue mass, expansion of the bone (except it must

be present in an A BC), sclerotic or nonsclerotic border (except it must be

nonsclerotic in giant cell tumor), presence or absence of bony struts or

compartments in the lesion, and size of the lesion.

If calcif ied matrix is identified in a lesion, it is tempting to narrow the

differential to either the osteoid series

or the chondroid series of lesions, depending on the character of the matrix. Be

careful of this. Very few radiologists can reliably differentiate chondroid fromosteoid matrix. Routine calcif ication of a lesion or debris, detritus, or

sequestrations in osteomyelitis can mimic ch ondroid or osteoid calcif ication and

be misleading. The only lesion that must exhibit calcif ied matrix is the

enchondroma (except in the phalanges). Chondroblastomas and osteoblastomas

demonstrate calcif ied matrix about half the time, and chondromyxoid fibromas

never have radiographically demonstrable calcif ied matrix.

P.1084

TABLE 41. Differential for ib Lesions

Fibrous dysplasia

Aneurysmal bone cyst


Enchondroma and eosinophilic granuloma

TABLE 41. Multiple Lesions FEEMHI




55/57

DIFFE ENTIAL DIAGNOSIS OF A SCLE OTIC

LESIONMany lytic lesions spontaneously regress and are not usually seen in patients

over 30 years of age. When these lesions regress, they often fill in with new

bone and have a sclerotic or blastic appearance. Therefore, when a sclerotic

focus is identified in a 20- to 40-year-old patient, especially if it is an

asymptomatic, incidental finding, the

following lesions should be considered: NOF (Fig. 41.41), EG, aneurysmal bone

cyst, solitary bone cyst, and chondroblastoma. Several other lesions should be

included that can also a ppear sclerotic: fibrous dysplasia, osteoid osteoma,

infection, brown tumor (healing), and perhaps a giant bone i sland (Fig. 41.42).

In any patient older than the age of 40 years, the number one possibility shouldbe metastatic disease.

Fibrous dysplasia

Eosinophilic granuloma

Enchondroma


Hyperparathyroidism (brown tumors)Infection

P.1085




56/57

FIGURE 41.41. Healing Nonossifying Fibroma. A plain film of the knee in

this 25-year-old patient reveals a sclerotic lesion in the proximal tibia,

which is a healing or resolving nonossifying fibroma.




57/57

EFE ENCES

1. David R, Oria R, Kumar R, et al. Radiologic features of eosinophilic

granuloma of bone. Pictorial essay. AJR Am J Roentgenol 1989;153:

10211026.

2. Dahlin D. Giant cell tumor of bone: highlights of 407 cases. AJR Am J

Roentgenol 1985;144:955960.

3. Gold R, Hawkins R, Katz R. Pictorial essay. Bacterial osteomyelitis:

findings on plain radiography, CT, MR, and scintigraphy. AJR Am J

Roentgenol 1991;157:365370.

FIGURE 41.4 . Giant Bone Island. A large sclerotic lesion is present in

the right supraacetabular region of the ilium (arro ), which represents a

giant bone island. The sli ghtly feathered margins of the trabeculae, which

blend in with the normal bone, and the long axis of the lesion being in the

direction of primary weight bearing are characteristic for a bone island.


Date post:	07-Apr-2018
Category:	Documents
Upload:	ranbir-mehmi
View:	226 times
Download:	0 times

Musculo Skeletal Radiology

Documents